Multiple Metal-Carbon Bonds. 13. Preparation and ... - ACS Publications

Journal of the American Chemical Society / lOl:12 / June 6, 1979. Multiple Metal-Carbon Bonds. 13. Preparation and. Characterization of Monocyclopenta...
0 downloads 0 Views 2MB Size
3210

Journal of the American Chemical Society

/

lOl:12

/ June 6, 1979

Multiple Metal-Carbon Bonds. 13. Preparation and Characterization of Monocyclopentadienyl Mononeopentylidene Complexes of Niobium and Tantalum Including the First Details of an a-Abstraction Process C. D. Wood, S. J. McLain,** and R. R. Schrock*2b Contribution from the Department of Chemistry, Massachusetts Institute of Technology, Cambridge, Massachusetts 02139. Received August 4, I978

Abstract: The neopentyl complexes, Ta(CH2CMe&X3 (X = CI or Br), react with TlCsH5 or LiC5Me5 to give the monocyclopentadienyl complexes, Ta(?S-C~Rs)(CH2CMe3)zX2 ( R = H or Me). Several of these decompose smoothly to give neopentane and the electron-deficient neopentylidene complexes, Ta(qS-CsR5)(CHCMe3)X2. I n Ta($-CSR5)(CHCMe3)X2 attraction of the CH, electrons by Ta cooperates with the steric effect of the tert-butyl substituent to give large M=C,-Cp angles (-165’) and low values for ’ J c H , (-75-85 Hz) and V C H , (-2500 cm-I). We propose that the electrophilic do metal in Ta(qsC ~ R S ) ( C H ~ C Msimilarly, ~ ~ ) ~ Xbut ~ more weakly, interacts with a CH, electron pair in one neopentyl ligand and thereby sets up H, for abstraction by the second neopentyl ligand. We show that ( I ) the a-abstraction reaction is intramolecular; (2) the rate of decomposition of TaCp(CHzCMe3)2Clz at 309 K varies with solvent in the order ether ( I ) < pentane (2) < benzene (4) < chloroform (15) dichloromethane; ( 3 ) the rate of decomposition of TaCp(CH2CMe3)2Brz is 200-400 times that of TaCp(CH2CMe3)2Cl2; (4) the rate of decomposition of a q5-C5Hs complex is 5 X IO3 times that of the corresponding q 5 - C ~ M ecomplex; s ( 5 ) the deuterium isotope effect in the rate-determining a-hydrogen abstraction step is approximately 6. Many of the results can be explained adequately if only cis-Ta(q5-C5R5)(CH2CMe3)2X~is prone to a-abstraction and transTa(q5-C5R5)(CH2CMe3)2X2 is inert. This postulate is fully supported by studies of the TaCp(CH2CMe3)(CH2Ph)X2 complexes where the cis/trans equilibrium is observable. The yields of analogous N b neopentylidene complexes are low. TaCp(CHCMe3)Clz reacts quantitatively at 25 “C with acetonitrile or benzonitrile to give imido complexes, TaCp[N(R)C=CHCMe3]C12, and with diphenylacetylene quantitatively at 25 “ C to give a new alkylidene complex, TaCp[C(Ph)(PhC=CHCMe3)]C12.

-

Introduction Since the discovery of the first primary alkylidene complex, Ta(CH2CMe3)3(CHCMe3),3 we have been trying to prepare one having no potentially reactive alkyl ligands and, like Ta(CHzCMe3)3(CHCMe3), less than 18 valence electrons. We have now prepared a class of Nb and Ta complexes of this nature with the formula M(q5-C5R5)(CHCMe3)X2 (R = H or Me; X = halide or alkyl). Since the precursors to several of the neopentylidene complexes can be isolated and characterized we can for the first time describe the &-hydrogen abstraction process in some detail including thermodynamic parameters, the deuterium isotope effect, and how the rate of a-abstraction depends on molecular structure, other ligands, and solvent. These results, as well as the preparation, the characterization, and a few simple reactions of such alkylidene complexes, are reported here.4 Results Reaction of Alkyl Complexes with Cyclopentadienyl Reagents. Ta(CH2CMe3)2C13 in toluene reacts rapidly with sparingly soluble TlCsH5. The first fleeting purple color changes to red within a few minutes. One mole of neopentane is ultimately given off and monomeric, red, crystalline, sublimable TaCp(CHCMe3)C12 (Cp = q5-C5H5) can be isolated from toluene/pentane (1:4) mixtures in 75% yield. Two byproducts of this reaction are sparingly soluble TaCpzClz and (if excess TICSHSis present) TaCp2(CHCMe3)C1.6In ether at 0 OC the same reaction rapidly gives purple TaCp(CH2CMe3)2C12 which only slowly evolves neopentane in this solvent (see next section) to give TaCp(CHCMe3)Cl2: Ta(CHzCMed2C13

TlCsHs

-

__*

TaCp(CH2CMe3)2C12

-CMe4

TaCp(CH2CMe3)2C12

can

TaCp(CHCMe3)Clz (1) also be prepared from

0002-1863/79/1501-3210$01 .OO/O

TaCp(CH2CMe3)C13 and 0.5 mol of Mg(CHzCMe3)2(dioxane) in ether. TaCp(CHzCMe3)C13 is prepared straightforwardly from Ta(CH2CMe3)C& and TlCsH5 in toluene. Ta(CHzCMe3)2Br3 reacts more rapidly than Ta(CH2CMe3)2C13 with TlCsH5 in ether or toluene to give TaCp(CHCMe3)Brz in high yield. The suggested intermediate, TaCp(CH2CMe3)2Br2, could not be isolated, but it could be detected by ‘H N M R and its decomposition followed in situ at low temperature (see later section). The reaction of Ta(CH2CMe3)(CH2Ph)X3 with TIC5H5 in toluene gives TaCp(CH2CMe3)(CH2Ph)X2in high yield (X = CI or Br). The bromide complex decomposes readily in solution at 25 OC to give TaCp(CHCMej)Br2 (see later section). The reaction between Nb(CH2CMe3)2CI3 and TlCsHs is comparatively complex. In toluene at 25 “ C most of the product is an insoluble, brown powder; purple-red NbCp(CHCMe3)C12 can be isolated easily but only in ca. 5% yield. At lower temperatures the reaction passes through a stage where TIC1 is evident before the brown precipitate forms. We have tried other solvents (ether, dichloromethane, THF) and other cyclopentadienyl reagents (MgCp2, NaCp) at -78 to 25 OC. Often (as with MgCp2 in ether) the reaction appears to proceed smoothly to a more orange compound (possibly NbCp(CH2CMe3)2C12) which cannot be isolated before it decomposes to the brown powder and traces of (if any) NbCp(CHCMe3)C12. Concomitant decomposition of Nb(CH2CMe3)2CI3 can also account in part for a low yield since, if more soluble, faster reacting TICSH4Me is used in toluene at 25 OC, NbCp’(CHCMe3)Clz (Cp’ = q5-C5H4Me) can be isolated in 30% yield. Neopentyl/halide complexes also react smoothly with LiCsMes in ether. Ta(CH2CMe3)C14, Ta(CH2CMe3)2C13, Ta(CH2CMe3)2Br3, and Ta(CH2CMe3)(CH2Ph)C13 give TaCp”(CH2CMe3)CI3, TaCp”(CH2CMe3)2C12, TaCp”(CHzCMe&Br2, and TaCp”(CH2CMe3)(CH2Ph)C12,respectively, in high yield (Cp” = q5-C~Me5).Only TaCp”0 1979 American Chemical Society

Wood, McLain, Schrock

/ Monocyclopentadienyl Mononeopentylidene Complexes

321 1 t

Keq CI-Ta-CH2CMe3 CI‘

,I

-

‘CH2Ph

C1-Ta-CH2CMej PhCHf

\Cl

I

3.0

Figure 1. The 270-MHz

a

5.0

7.0

(T)

9.0

’HNMR spectrum of TaCp(CH2CMe3)(CH2Ph)CI2at -40 “C in CDCI3: c = cis form; t

= trans form.

Table 1. Thermodynamic Data for the Cis/Trans Interconversion in TaCp(CHzCMe3)(CH2Ph)Xz comDd TaCp(CHzCMe3)(CHzPh)Cl2

solvent

T. K

Kea(t/c)

ether

25 1 23 I 203 333 238 223 201 233 218 243 232 219 333

4.4 2.1 1.3 46 1.5 1.3 1.1 0.7 0.6 0.15 0.25 0.33 0.02d

toluene CHC13 TaCp(CHzCMe3)(CH2Ph)Br2

ether

A H . kcal/mola 2.52

AS, eub

AG,* kcal/molc

13.3

-0.9

-4.1

-1.0

-3.8

-3.48

13.6 f 0.2

-18.1

13.2 f 0.2

A S is the entropy of the trans form minus that of the cis form. For a AH is the enthalpy of the trans form minus that of the cis form. the interconversion of cis and trans forms at -260 K (the coalescence point for the two C p signals in ether). Calculated using the determined values for AH and AS.

(CH2CMe3)2Br2 decomposes smoothly in chloroform at 3 1 OC to give TaCp”(CHCMe3)Brz (vide infra). Reaction of Nb(CH2CMe3)2C13 with LiCsMes so far has not given an identifiable product. ‘H NMR Data for Dialkyl Complexes. The IH N M R spectrum of TaCp(CH2CMe3)(CH2Ph)C12 in CDC13 at 270 M H z and -40 OC is shown in Figure 1. We assign one set of signals to a trans form of tetragonal pyramidal TaCp(CH2CMe3)(CH2Ph)Clz based on the fact that there is one sharp resonance for the neopentyl a! protons and one sharp resonance for the benzyl a! protons (HJ. (The crystal structure of TaCp”(PMe3)2(CPh)C17shows it to be an almost perfect tetragonal pyramid so at this stage any trigonal bipyramidal geometry seems unlikely.) The other set we assign to cis-TaCp(CH2CMe3)(CH2Ph)C12based on the fact that each set of (Y protons now forms a diastereotopic pair which gives rise to a typical AB pattern (HcHL). As the temperature of the sample is raised toward 0 OC the two isomers begin interconverting. At 260 K in ether AG* = 13.6 f 0.2 kcal mol-’ (from coalescence of the two C p peaks). Note that the cis and trans forms interconvert at a rate ( k = 15 s-l) which is extremely

s-I) at which a fast compared to a typical rate ( k = compound of this type decomposes (see later sections). The equilibrium between the cis and trans forms (eq 2)

Fp



C1-Ta-CH2CMe3 ’CHJ’h

K,,(tlc)

F”



Cl-Ta-CCHgMe,

(2)

PhCH, a ’ varies with temperature and solvent as shown in Table I. We should first note that more of the cis form (which should have a higher dipole moment than the trans form) is found in chloroform (dielectric constant E = 4.8 at 20 “C) than in toluene (€200~ = 2.2). Ether ( E Z O O C = 4.3) therefore seems out of place but it is also the only one of the three solvents which could coordinate to the metal. Secondly, K , increases (giving more trans form) as the temperature increases. From these data one can conclude that AHtlc and AS,!, vary markedly with solvent. In the two relatively noncoordinating solvents both A H and A S are small compared to their values in ether. The ‘ H N M R spectrum of TaCp(CH2CMe3)(CH2Ph)Br2 C1

1 101:12

Journal of the American Chemical Society

3212

June 6 , 1979

2’1 2.0

t:

-8.5 In -

-9.01

1

L

t

I

(io 1

M I N INTERVALS) I 1

1

I

I

Figure 2. First-order plots of the decomposition of (a) TaCp(CD2CMe3)2C12in CDC13 at 309 K; (b) TaCp(CHzCMe3)2C12 in CDC13 at 280 K.

in ether at low temperatures is similar to that of TaCp(CH2CMe3)(CH2Ph)C12 and AG* = 13.2 f 0.2 kcal mol-’ for interconverting the two forms (Table I). The main difference is that the cis form is more favored as the temperature increases ( K e gdecreases). Therefore at 60 “ C in ether K,, for TaCp(CHzCMe3)(CHzTh)Cl2 is 46 (essentially all trans form) while Kep for TaCp(CHzCMe3)(CH2Ph)Br2 is 0.02 (essentially all CIS form). In contrast to the spectra of TaCp(CH2CMe3)(CH2Ph)X2, the IH N M R spectrum of TaCp”(CH2CMe3)(CH2Ph)C12 shows no evidence for a cis form between -80 and 80 “C. (We believe it unlikely that the cis and trans forms are still interconverting rapidly at -80 “C.) Keq(t/c) must be on the order of 100 or more so that AG,,, is at least 3 kcal mol-’. The most reasonable explanation is that the more sterically demanding v5-C~Me5group “locks” the alkyls trans. The little cis form which is present would still be forming rapidly from the trans form on the chemical time scale since we have found AG* for pseudorotation in Cpf’ metallacyclopentane complexes of this type (a process which may be closely related to the cis-trans isomerization) to be only 2-3 kcal mol-’ higher than in C p metallacyclopentane complexes.8a Turning now to the Cp and Cp” dineopentyl complexes we find that cis forms again cannot be observed directly. The reason is clearly steric in the case of TaCp”(CHzCMe&C12 (cf. TaCp”(CH2CMe3)(CH2Ph)CI2 above). It is probably steric in origin for TaCp(CH2CMe3)2C12also; a neopentyl and benzyl group can be mutually cis but two neopentyl groups would prefer to be mutually trans. However, we cannot say that no ci~-TaCp(CHzCMe3)2C12is present at 25 “C; as much as 5% could be present and go undetected by ‘ H N M R . Finally, the I3C N M R spectrum (gated decoupled) of TaCp(CH2CMe3)(CH2Ph)CIz could be telling us something about the degree to which a-abstraction is likely in a neopentyl ligand vs. a benzyl ligand. In the cis form the benzyl a-carbon resonance is a clean 1:2:1 triplet ( ‘ J C H , = 125 Hz) which suggests that I J C H ,is the same as JcH,/ (Ha and Ha’ are diastereotopic). However, what we can see of the neopentyl a-carbon resonance (it overlaps with the cyclopentadienyl resonances) is similar to that found in TaCp(CHCMe3)(CH2CMe3)Cl (for example) where ‘JcH, = 100 and ‘JcH; = 110 Hz (see later). We conclude that whatever is causing

I

I

314

33

316

Figure 3. Plot of In ( k / T ) vs. I / T for decompositions of TaCp(CHzCMe3)2C12 in CDC13 at five temperatures (see Table 11).

low values for JcH,, it is more likely to operate in a neopentyl than in a benzyl ligand. Details of the Thermal Decomposition of TaCp(CH2CMe&Xz, X = C1 or Br. The decomposition of TaCp(CH2CMe3)2C12 in the dark is first order through at least 4 half-lives in ether, pentane, benzene, chloroform (Figure 2), and dichloromethane at several temperatures (Table 11). The decomposition product when the reaction is rapid in the dark at 95% TaCp(CHCMe3)Cll (which is stable under the reaction conditions). However, at higher temperatures, signals which correspond to TaCp(CHzCMe3)C13 and TaCp(CHCMe3)(CH2CMe3)CI (vide infra) in about equal amounts, as well as some unidentified signals, are found in the ] H N M R spectrum. These byproducts amount to 10-20% of the product mixture in a slow reaction (e.g., at 3 15 K in ether) but only a few percent in a fast reaction (e.g., at 299 K in chloroform). In the few cases where significant byproducts formed k was determined from measurements made early in the reaction. However, even over longer periods (1 -2 half-lives) their production did not markedly affect the linearity of the first-order plot for appearance of TaCp(CHCMe3)C12. Plots of In ( k / T )vs. 1/ T (see Figure 3 for an example) gave values for AH+ and AS* for the decomposition of TaCp(CH2CMe3)2C12 in three solvents, and for decomposition of TaCp(CH2CMe&Br2 (less accurately) in one (Table 111). By comparing the calculated rates in these three solvents at 309 K (Table 11) with that in pentane at 309 K we obtain the ordering ether ( ~ 1 < ) pentane (1.7) < benzene (4.1) < chloroform (14.8). At 268 K the rate in dichloromethane is about the same as that in chloroform. If AS* in dichloromethane is similar to that in chloroform, they should also be about the same at 309 K. Therefore at 309 K the ordering is roughly ether ( ~ 1 < ) pentane (-2) < benzene (-4) < chloroform dichloromethane (-1 5). Of course if for two solvents are quite different, the relative rates will change markedly with temperature. For example, at 294 K the rate in chloroform is 30 times that in ether (Table 11). In the presence of 1 mol of Et4N+CI- TaCp(CH2CMe3)2C12 decomposed at 268 K in dichloromethane with k = 0.012 (fO.OO1) min-I, the same result as in the absence of Et4N+CI- (Table 11). In the presence of 5 mol of EtdN+CIthe rate was again unchanged. Therefore, loss of C1- from the

as*

-

Wood, McLain, Schrock

/ Monocyclopentadienyl Mononeopentylidene Complexes

3213

Table 11. Kinetic Data for Decomposition of Dineopentyl Cyclopentadienyl Complexes"

compd

solvent ether

pentane benzene

chloroform

dichloromethane chloroform ether TaCp"(CHzCMe3)2Br2 ~

_

_

_

_

_

_

~

~

~~~

~

_

chloroform dichloromethane _ _ ~ ~ ~

T, K

k X IO3, min-'

328 322 315 309 308 304 303 294 27 5 262 309 309 304 302 29 5 309 304 299 294 289 285 28 I 268 268 309 304 275 262 304 304

81 f 3 38f I 24 f 1 13 h 12.0 f 0.5 7.56 6.3 f 0.5 2.5b 0.263b 0.046b 22f 1

t ~ pmin , 8.5f 0.5 18.0 f 0.5 296 1

57 f 2 110f4

32 f 5

536

29f I 24f 1 9.1 f 0.5 1936 1426 100 f I O 77 f 5 52 f 2 38 f 2 31 f 1 126 11 f 1 36f 1 1778 100 f I O C 22 f 5 c 6.0f 0.5 6.3f 0.5

24 f 1 29 f 1 71 f 2 7fl 9 f l 13f 1 1851 22f I

63 f 5 20f 1 7fl 32 f 4 115 f I O IO9 f I O

~

Data were obtained by IH N M R integration of C p or Cp" resonances in starting material vs. product vs. time through at least 3 half-lives. The rate constant was determined by a least-squares fit of the data; correlation coefficients were always >0.98. See Figure 2 for a plotted example. Errors in k were determined by a standard statistical method based on standard deviations. The exception is TaCp(CH2CMe3)2Br2.' Calculated from AH* and AS* given in Table 111. C T h erate constant was determined by estimating the half-life by IH N M R for decomposition of TaCp(CH2CMe&Br2 in a mixture of TaCp(CHzCMe3)2Br2 and TaCp(CHCMe3)Brz (see Experimental Section). a

Table 111. Thermodynamic Data for Decomposition of T a C o ( C H K M e 3 h X 7 ( X = CI or Br)"

in the TaCp(CHCMe3)Clz present in the TaCp(CDCMe3)Cl2/TaCp(CHCMe3)Cl2 mixture vs. the Cp signal suggests that kH/kD = 7 f 1 for the a-abstraction step. We conclude X solvent AH*. kcal mol-' 6 AS*. eub that the rate-determining step involves breaking a carbon-H, CI ether 19 f l(18.74) -16 f 4 (-15.3) (or D,) bond and kH/kD = 6. (The possibility of a prior CI benzene 21 f 2 (21.28) -4 f I O (-4.2) equilibrium will be discussed later.) CI chloroform 10.7 f 0.5 (10.71) -36 f 2 (-35.8) A sample of TaCp(CD2CMe3)2C12 in benzene decomposed Br ether 16 f 2 (16.14) -13 f 7 (-12.9) essentially completely at 60 OC in 4 h to give a mixture of " AH* and AS* were determined by a least-squares fit of In ( k / T ) neopentane-d3 (94%) and neopentane-d2 (6%). Up to 2-3% vs. 1 / T using the data in Table I1 (see Figure 3 for an example). The of the neopentane-d2 probably arises from -CHDCMe3 groups errors were determined by a standard statistical method based on (since kH/kD 6) but that remaining (3-4%) must arise in standard deviations except for TaCp(CH2CMe3)2Brz (only two some other manner such as formation of free radicals (see points), for which extremes were calculated based on the error in tll2 later), y-abstraction, cleavage of a cyclopentadienyl C-H a t each temperature. The errors in AS* and AH* are the experibond, etc. We do not believe that this anomaly is serious mental errors. Real errors, especially for AS*, may be somewhat enough to prevent our concluding that only an a-deuterium larger owing to the relatively small temperature range (-25 "C) in atom is abstracted in this case, or, in general, only an a - h y which rates were measured. Thevalues in parentheses are the exact drogen atom is abstracted when TaCp(CHzCMe3)2C12 devalues obtained from the least-squares plot of In ( k / T ) vs. 1 / T . Exact composes. values must be used to calculate k a t T. We can show that the a-abstraction reaction is intramolecular in the following way. A mixture of TaCp(CH2CMe3)2C12 (25 mg) and TaCp(CD2CMe3)2C12 (150 mg; this tantalum complex cannot play a significant role in its decomratio negates any intermolecular isotope effect on the order of position to the neopentylidene complex. 6) in CHC13 at -22 "C after -0.5 h decomposed partially to TaCp(CD2CMe3)2Clz decomposes significantly more slowly give a mixture of 41% d3,7% d2,2% d l , and 50% do neopenthan TaCp(CH2CMe3)2Cl2. I n chloroform the rate constant tane. All volatiles were removed and the remaining dineopentyl a t 309 K is 0.036 min-l; that calculated for complex (mostly TaCp(CDzCMe3)2CI2) was redissolved in TaCp(CH2CMe3)2C12 is 0.193 min-', or kH/kD = 5.4 (f0.5). CHC13; it decomposed essentially completely in the next 5 h This isotope effect can also be measured in a competitive exto give a mixture of 61% d3, 16% d2,5% d l , and 18% do neoperiment. The product of the reaction of Ta(CHDCMe3)2C13 pentane. The formation of almost exclusively neopentane-d3 and TlCsHs in toluene is largely TaCp(CDCMe3)Clz. A and -do in the early part of the decomposition confirms that careful IH N M R integration of the neopentylidene a proton the a-abstraction reaction is intramolecular. Formation of ~~

3214

Journal of the American Chemical Society

/ I01:12 / June 6, I979

much more neopentane-d2 and -dl in the latter part of the is an oil under these conditions (cf. TaCp”(CHCMe3)Bra decomposition is most likely due either to intermolecular alkyl below). We believed that retardation of the a-abstraction reaction exchange prior to intramolecular a-abstraction or to intermolecular a-abstraction. We cannot implicate one or the other on replacing Cp with Cp” could be counteracted by replacing based on present data. CI with Br. Indeed TaCp”(CH2CMe3)2Br2 decomposes significantly more rapidly than TaCp”(CH2CMe3)2CIz. In The reaction between Ta(CH*CMe3)2Br3 and TIC5H5 in ether at -40 “ C gives a mixture of TaCp(CH2CMe3)2Br2 and chloroform or dichloromethane at 3 1 OC its half-life is ca. 1 I O TaCp(CHCMe3)Brz (by ‘ H N M R ) . Conversion of the former min and k = 0.006 min-l (Table 11). Similar data for Tato the latter was followed by ‘ H N M R at 2 and -11 OC. Cp”(CH2CMe3)2C12 are unavailable for comparison but we Comparing these rate constants (Table 11) with those calcu- might expect the rate of decomposition of TaCp”(CH2Clated for the decomposition of TaCp(CH2CMe3)2Cl2 at 275 Me3)2C12 to be about 240 times slower than that of TaCp”and 262 K (Table 11) gives kBr/kC1 = 380 and 480. (At 304 (CH2CMe3)2Br2 at 304 K by comparison with the q5-C5H5 K the calculated ratio is about 240.) Therefore the bromide system above. If so we can calculate the consequence of recomplex decomposes about 400 times more rapidly than the placing a Cp” by a Cp ligand: [k(TaCp(CHzCMe3)2Clz, chloroform, 304 K ) X 240]/[k(TaCp”(CHzCMej)2Br2, chloride complex in ether at ca. 0 OC. An isotope effect was obtained (as above) by measuring how much chloroform, 304 K] = 5680. This can also be calculated by extrapolating the TaCp(CHzCMe3);?Br2data in ether to 304 TaCp(CHCMe3)Brz formed on decomposition of TaCp(CHDCMe3)2Br2. The result, kH/kD = 6 f 1, does not K and assuming that k(chloroform)/k(ether) = 19 (the ratio differ measurably from that obtained above for the corre- of rates for the chloride complex at 304 K) to give k C p / k C p , , = 5630). Therefore the rate of a-abstraction increases by 5 X sponding chloride complex. Finally, we should note that the intramolecular a-abstracl o 3 on replacing a q5-CSMe5 ligand by a q5-C5H5 ligand. tion step also proceeds smoothly in the solid state. If a sample TaCp”(CHCMe3)Brz can be recovered easily from a typical decomposition in chloroform and recrystallized from pentane of TaCp(CH2CMe3)2C12 is warmed to 70 OC in a sublimer in (in which it is very soluble) or sublimed at 110 OC (it inelts vacuo, TaCp(CHCMe3)Clz rapidly sublimes onto the waterbefore subliming at 1 p). cooled probe (94% yield). All the a-abstraction reactions we have mentioned are acDecomposition of Other Dialkyl Complexes and the Role of Light. TaCp(CH2CMe3)(CH2Ph)Cl2 decomposes smoothly celerated markedly by light and their sensitivity to light paron heating in vacuo to 70 O C . Toluene (68%) and neopentane allels the ease of the thermal reaction. For example, (10%) are found in the liquid N2 trap (by IH N M R vs. internal TaCp”(CH2CMe3)zClz must be photolyzed in benzene in a quartz or Vycor tube using 360-nm high-intensity light from standard) and TaCp(CHCMe3)Clf (60% yield) is found on the water-cooled probe. In C6D6 at 60 “C TaCp(CHzCMe3)a medium-pressure Hg lamp. On the other hand, a benzene (CH2Ph)CIl decomposes similarly (57% TaCp(CHCMe3)C12, solution of TaCp”(CH2CMe3)lBr2 (50 mg in 1 mL) turns cranberry red in bright indirect sunlight in 15-30 min; by IH 59% toluene, 29% neopentane) with a half-life of about 35 min N M R the sample is mostly TaCp”(CHCMe3)Brz. Finally, (k = 0.020 min-I). We cannot measure the rate accurately since the reaction is not straightforward. In neither case have solutions of TaCp(CH2CMe3)2Clz (the isolable compound we been able to demonstrate that TaCp(CHPh)Clz is formed; most easily converted into TaCp(CHCMe3)Clz thermally) are sensitive even to room fluorescent light. In fact the preparative this species is still unknown. yield of TaCp(CHCMe3)C12 in toluene can be improved sigTaCp(CH2CMe3)(CHzPh)Brz decomposes in ether at 25 nificantly by covering the flask with aluminum foil. “ C much more cleanly than TaCp(CHzCMe3)(CH2Ph)CI2. The photochemical reaction is no cleaner than the correOnly about 10%neopentane forms and TaCp(CHCMe3)Br2 sponding thermal reaction, especially if relatively concentrated can be isolated in high yield. For this decomposition at 305 K samples are used. The most important difference is that the k = 0.029 min-l. If we assume that the rate will approximately photochemical reaction produces significant quantities of double every 10 “ C then k at 60 “ C will be about 0.23 min-I. The rate of decomposition of TaCp(CH2CMe3)(CH2Ph)C12 2,2,5,5-tetramethylhexane (dineopentyl), possibly by coupling in ether at 60 “ C should be l/4 that in benzene or k = 0.005 of neopentyl radicals. Therefore it does seem more complex min-I. The bromide complex in this case therefore would de- than the thermal reaction and its use as a preparative technique compose about 50 times faster than the chloride complex at therefore somewhat limited. We want to stress that these 60 OC. This correlates with the relative amounts of the cis photochemical results are preliminary. More careful, quantiforms for each; TaCp(CH2CMe3)(CH2Ph)Br2 is essentially fiable studies of the photolysis at low conversion are clearly all cis while TaCp(CH2CMe3)(CH2Ph)Cl2 is only -2% cis needed in order to understand exactly how light influences the a-abstraction step. or the ratio of cis bromide to cis chloride is about 50. (Note that Other Neopentylidene Complexes. TaCp(CHCMe3)Clf the relative rates of decomposition of TaCp(CHzCMe3)2Br2 and TaCp(CH2CMe3)2C12 (vide supra) would be on the order reacts smoothly with first 1 mol of LiCH2CMe3 in pentane to of 100 at about 60 “C.) Finally, the rate of decomposition of give TaCp(CHCMe3)(CH2CMe3)C1, then a second to give TaCp(CHCMe3)(CH2CMe3)2. Both products are extremely TaCp(CH2CMe3)(CH2Ph)Brz in chloroform ( k = 0.155 min-I at 305 K) is only about five times what it is in ether. (It soluble in pentane and for that reason difficult to isolate on a small scale by crystallization at low temperature. At room is essentially all cis in each solvent.) Both Ta C p”( C H 2C M e 3) ( C H 2Ph) C 12 and T a Cp”- temperature, in fact, they are both sublimable, yellow oils. The (CHzCMe3)2C12 are fairly stable thermally. They decompose analogous complexes obtained by replacing one or two chloride slowly in C6D6 at 70 OC (days), but, like the forced decom- ligands in TaCp(CHCMe3)Clz with methyl groups are also positions of TaCp(CH2CMe3)2Clz, the reaction does not give extremely pentane-soluble oils (possibly low-melting solids when pure) which we cannot purify beyond 90-95% by crysTaCp”(CHCMe3)C12 cleanly. Signals due to TaCp”(CHzCMe3)C13 can be seen in the ‘H N M R spectrum in each tallization or sublimation. case, and in the former also signals due to T a c ~ ’ ’ ( C H P h ) c 1 2 . ~ ~ Two analogous v5-C5Me5 complexes are more tractable and I n neither case, however, is Me3CCHzCH2CMe3, have been fully characterized. Orange TaCp”(CHCMe3)(CH2CMe3)CI and yellow TaCp”(CHCMe3)(CH2CMe3)2 PhCHzCH2Ph, or PhCH2CH2CMe3 a product. TaCp”(CHCMe3)CIz can be isolated >90% pure by tedious fractional can be prepared by treating TaCp”(CH2CMe3)C13 with 2 and sublimation of the mixture resulting from thermal decompo- 3 mol of LiCH2CMe3, respectively, in pentane. They are both sition of TaCp”(CH2CMe3)(CH*Ph)C12 in the solid state. It extremely soluble in pentane but can be isolated crystalline

Wood, McLain, Schrock

/

Monocyclopentadienyl Mononeopentylidene Complexes

3215

Table IV. Pertinent ‘ H , 13C N M R , and IR Data for Neopentylidene Complexes”

compd

IHm,7

I3Ca, ppm

’JcH,, Hz

TaCp(CHCMe3)Clz TaCp(CHCMe3)(CHzCMe3)cl TaCp(CHCMe3)(CHzCMe3)z TaCp”(CHCMe3)Clz TaCp”(CHCMe,)(CHzCMe3)cI TaCp”(CHCMe3)(CHzCMe3)z TaCp(CHCMe3)Brz TaCp”(CHCMe3)Brz TaCp(CHCMe3)MeCI TaCp(CHCMe3)Mez TaCp(CHCMe,)CIz( P M e d N bCp’( C H C Me3)CIz N bCp(CHCMe3)C12

3.62 4.08 5.01 5.13 6.27 6.87 3.56 5.18 4.12 4.99 4.01 2.70 2.75

246 240 234 242 232 228 254 249

84 81 80 83 76 76 83 77

2510 (2490)c,d 2520 (2490) 2515 (2500) 2435 2460 2440 2470 (2510)d 2420

23 1 272 254

78 84 95f

2475 (2510) 25OOc (2475)‘ 2535p 2520

vCHa,

cm-1

b

a Cp,= q5-C5H5, Cp’ = qS-CsHdMe, Cp” = v5-C5Me5. Full details can be found in the Experimental Section. N M R solvent = C6D6 unless otherwise noted. I n Nujol. The minor peak or shoulder, if any, is at the frequency shown in parentheses. V C D ~= 1855 cm-I in TaCp(CDCMe3)Clz. See Figure 4. e Comparatively weak. f Broad due to coupling of C, to 93Nb ( I = 912, 100% abundant).

from concentrated solutions at -30 OC. TaCp”(CHCMe3)(CH2CMe3)C1 sublimes at 60-70 OC and TaCp”(CHCMe3)(CH2CMe3)2 at 100 OC (1 F ) without decomposition. The reaction of TaCp(CHCMe3)Cl2 with LiCD2CMe3 gives TaCp(CHCMe3)(CD2CMe3)CI; no deuterium transfers from the neopentyl to the neopentylidene a-carbon atom in 16 h at 70 “C. TaCp(CDCMe3)(CH2CMe3)2 (prepared from TaCp(CDCMe3)Cl2 and 2LiCH2CMe3) is also similarly stable toward H transfer. Apparently, therefore, a-hydrogen atom scrambling is slower in these species than in Ta(CHzCMe3)3(CDCMe3) where LG* 28 kcal This finding suggests that we can measure the isotope effect competitively for a reaction involving a lithium reagent. This may be important since k H / k D for formation of Ta(CH2CMe3)3(CHCMe3) (in the presence of LiCH2CMe3) was found to be -3, not -6 as found here for decomposition of T ~ C P ( C H ~ C M ~ TaCp”(CHDCMe3)2C12 ~)~X~.~~ was treated with 2 mol of LiCHDCMe3 in pentane and the resulting mixture of TaCp”(CHCMe3)(CHDCMe3)2 and TaCp”(CDCMe3)(CHDCMe3)2 examined by 270-MHz ‘H N M R . Integration of =CHCMe3 vs. all Cp” protons allowed us to calculate kH/kD = 6 f 1 for this reaction. Almost certainly, therefore, the lithium plays no role similar to that it plays in the formation of T ~ ( C H Z C M ~ ~ ) ~ ( C Hand CM TaCp”(CHCMe3)(CH2CMe3)CI (the initial all-protio product) most likely forms by intramolecular decomposition of TaCp”(CH2CMe3)3CI. NMR and IR Data for Neopentylidene Complexes. Pertinent N M R and IR data for neopentylidene complexes are given in Table IV. The chemical shift of the neopentylidene 01 proton varies from about T 3 to about T 7 . This could be ascribed largely to an electronic effect, Le., the a-proton resonance moves to higher field on replacing a chloride with a more electrondonating alkyl ligand. The same happens on replacing Cp with the more electron-donating9 Cp” group. All the H, chemical shifts are at higher field than in the biscyclopentadienyl alkylidene complexes of the type MCp2(CHR)X6 ( 7 -2 to 0). The only other notable feature of the IH N M R spectra of the neopentylidene complexes is the fact that the neopentyl cy protons in TaCp(CHCMe3)(CH2CMe3)CI and TaCp(CHCMe3)(CH2CMe3)2 (and the Cp” analogues) are diastereotopic, and differ in chemical shift by up to 2 ppm. In TaCp(CHCMe3)(CHzCMe3)2, for example, one set is found at T 8.44, the other at T 10.52 with 2 J ~ , H =B 12 Hz. Therefore we can at least say that these molecules are stereochemically rigid, pseudotetrahedral species on the ‘H N M R time scale. The I3C N M R spectra of the neopentylidene complexes all

show a resonance for the neopentylidene cr-carbon atom within what is turning out to be the “normal” range for nucleophilic alkylidene complexes, 224-275 ppm downfield of Me& (Table IV). The peak shifts upfield (as does the H, peak in the IH N M R spectrum) on successively replacing the two chloride ligands by alkyl ligands or Cp by Cp”. It shifts downfield on replacing C1 by Br. I J c H , is surprisingly and consistently lower than what is found in analogous (and more crowded) 18electron complexes such as TaCp2(CHCMe3)CI (121 Hz) or NbCp2(CHCMe3)CI (131 H z ) . The ~ change in ‘ J C H , o n replacing CI by Br or Cp by Cp” is less than that on replacing Ta by Nb; I J c H , increases by about 10 Hz (cf. 121 to 131 H z above). It should be pointed out that I J c H , for any neopentyl ligand also seems significantly lower than it should be (ca. 125 Hz) for an sp3-type C-H bond, and J C H , # J c H ~ in; every case we estimate J C H , = 100 f 2 and J C H ~= 1 I O f 2 Hz. J C H ,in the methyl group bound to Ta in TaCp(CHCMe3)Me2 is also slightly lower (1 18 Hz) than normal (125 Hz). The IR spectrum of TaCp(CHCMe3)Cl2 in Nujol is shown in Figure 4a. The peak we want to draw attention to most is at 2510 cm-I; it has a weak shoulder at 2490 cm-I. The IR spectrum of TaCp(CDCMe3)Clz (Figure 4b) shows a similar peak at 1855 cm-I, a shift of the former by 111.35. We therefore assign the 25 IO-cm-I peak to the C-H, stretch. ~Other ~ ) , ~peaks ~ in the spectrum of TaCp(CHCMe3)Cl2 change on replacing H, by D, the most prominent being those at 1255 and 1168 cm-I, which appear to shift up (to 1285 cm-l) and down (to 1125 cm-I), respectively. At this juncture, however, it would be premature to attempt to assign any of these lower frequency bands or attempt to rationalize why they shift. All are present (with minor changes in intensities and location) in the IR spectrum of TaCp(CHCMe3)Cl2 in benzene. Similar low V C H , bands are found in the IR spectra of other neopentylidene complexes (Table IV). The band is strongest and sharpest at the lowest frequency ( U C H , in TaCp”(CHCMe3)Br2, the lowest, is one of the strongest bands (all medium strength) in the spectrum) and weaker and broader at higher frequencies. Curiously U C H , is sometimes doubled, both in Nujol (see Figure 4c for TaCp(CHCMe3)Brz) and in solution (Figure 4d). The spectrum of TaCp”(CHCMe3)Brz again shows only a single, sharper peak. The IR spectrum of TaCp2(CHCMe3)CI6 in Nujol shows no low-frequency V C H , band. In the spectrum of TaCp2(CDCMe3)CI V C D , is clearly seen at 2150 cm-l. This suggests that V C H , occurs at 2900 cm-I in TaCp*(CHCMe3)C1, assuming that VCH/VCD = 1.35 as found above in the monocyclopentadienyl complexes. This does seem slightly low for an “olefinic” C-H stretch; it is more consistent with a C-H bond having a higher percentage p character, e.g., an aliphatic

3216

Journal of the American Chemical Society

/

I

c1

/ June 6, 1979

I

CP

J’ 4

101:12

PhCECPh V

CHCMe3

(r or r’ = H or CMe,) 1

w I

,

,

,

,

,

,

,

,

~

I

I

,

>

,



~

~



;



1

1000

2000

,000

!

Figure 4. The I R spectra of (a) TaCp(CHCMe3)Cll in Nujol; (b) TaCp(CDCMe3)C12 in Nujol(0 denotes the bands which are not common to both spectra); (c) TaCp(CHCMe3)Brz in Nujol; (d) TaCp(C HC Me3) Br2 in benzene.

-

Ph

4Ta,

Ph‘CHCMe3

excess^,^,

m

CpC1,TaCH,CHPCH2CHL

n

L

+ CH2=CHC\HPh

/C=CHCMe3

Ph C-H stretch. The main point, however, is that the V C H band ~ is not nearly so “abnormal” in TaCpz(CHCMe3)Cl as in TaCp(CHCMe3)C12. Reaction of TaCp(CHCMe3)CIz with Nitriles, PMe3, and PhCeCPh. TaCp(CHCMe#212 reacts at 25 “ C with acetonitrile or benzonitrile to give (quantitatively) orange, crystalline complexes whose IR, ‘ H , and 13C N M R spectra and elemental analyses are entirely consistent with their being a mixture of the E and Z isomers of the imido complexes formed by inserting RCN into the Ta=CHCMe3 bond3=(eq 3). This

P A

CpCl,Ta=CHCMe,

+ FEN

-

/

N

4

CpClzTa=C

/H ‘CMe3

R = Me or Ph is consistent with the neopentylidene ligand behaving as a nucleophile. But since 18 valence electron complexes such as TaCp2(CHCMe3)C16 or TaCp2(CH2)Me10do not react with acetonitrile readily we postulate that RCN must first coordinate to Ta before it is attacked by the nucleophilic neopentylidene ligand. This seems reasonable since the metal is electron deficient (14 valence electrons) and the coordination sphere is not especially crowded. Good evidence that this is the case consists of formation of a PMe3 adduct. TaCp(CHCMe3)Clz(PMe3) contains a neopentylidene ligand in which J C H ,and V C H , are virtually the same as in the starting material; the only noteworthy change is a shift of 613C, from 246 to 273 ppm. We therefore believe that PMe3 adds to the metal and is not in any way associated with Ha or C, of the neopentylidene ligand. Trimethylphosphine most likely takes up the fourth position in the tetragonal plane of a pseudo-tetragonal-pyramidal molecule, but we do not know if it is cis or trans to the neopentylidene ligand. The reaction of TaCp(CHCMe3)C12 with PhC=CPh is analogous to that with R C N (eq 4). Diphenylacetylene probably first coordinates to the metal and is then attacked by the neopentylidene a-carbon atom to give a metallacyclobutene intermediate (1) which rearranges to the observed product, blue 2, which by ’ H and 13CN M R is a single isomer. (The a-car‘CMe,

3

bon atom resonance in 2 is a singlet in the gated decoupled I3C N M R spectrum at 259 ppm.) Apparently a second “insertion” is unfavorable for steric reasons and/or because the a-phenyl group mediates the alkylidene’s nucleophilicity. TaCp(CPhPhC=CHCMe3)C12 (like TaCp(CHCMe3)C125) reacts smoothly with ethylene to give C12CpTaCH2CH2CH2CH2 and a single organic product which by 13CN M R is E- or 2 - 3 (eq 4). This is the same type of cleavage product (a terminal olefin) as that formed on reaction of TaCp(CHCMe3)Clz and ethylene. Internal olefins can be the major cleavage product in other types of alkylidene complexes (e.g., M(CHCMe3)C13(PMe3)2, M = N b or Tall). It is somewhat surprising, therefore, that no conjugated diene is formed. The reaction of TaCp(CHCMe3)Clz with dimethylacetylene is comparatively complex. The initial product is formed from 2 mol of dimethylacetylene per Ta, even at -78 OC. The neopentylidene a-carbon resonance disappears from the I3C spectrum and no other alkylidene a-carbon resonance appears. At low temperature five different methyl groups can be seen. On warming, the sample becomes paramagnetic. Details concerning this reaction and reactions of neopentylidene complexes with other acetylenes will be published separately. However, we might speculate at this time that a simple insertion reaction such as that found above may be the exception rather than the rule. Discussion We will first discuss the curious spectroscopic and structural features of the neopentylidene complexes reported here (and compare them with others) and then discuss how and why these complexes form; we believe that the former bears an important relationship to the latter. The reactions of TaCp(CHCMe3)C12 with acetonitrile and benzonitrile are similar to those of Ta(CHzCMe3)3(CHCMe3)3aand deserve no special comment at this time. The reaction of TaCp(CHCMe3)C12 with PhC=CPh is interesting12 but we want to put off any discussion until more examples (in this or another system) are available. The low values for IJCH, and V C H , suggest that TaCp(CHCMe3)C12 and analogous molecules differ significantly in some fundamental way from the more crowded, 18 valence electron complex, TaCpz(CHCMe3)CI, whose structure is known,13 and for which ‘JcH,and V C H , are not SO unusual [121 Hz and 2900 cm-l (VCD X 1.35, vide supra), respectively]. Structures for two other molecules in this class are known ( T ~ C P ~ ( C H P ~ ) ( C H and ~ PTaCp2(CH2)Me14). ~)~ Interestingly, l J c H , decreases from 132 Hz in TaCps(CH2)Me to 126 Hz in TaCp2(CHPh)(CH2Ph) to 121

Wood, McLain, Schrock

/ Monocyclopentadienyl Mononeopentylidene Complexes

H z in TaCpz(CHCMe3)CI, while the M=C,-R angle (e) increases from 125’ ( R = H ) to 135’ (R = Ph) to 150’ ( R = CMe3), W e might therefore postulate an inverse correlation of IJcH, with 8. This seems reasonable since as 8 increases (for R = Ph and CMe3) Ha would necessarily be pushed into a more and more unusual position. When 6 = 180’ (an extreme which will illustrate the point) the C-H, bond could be described in valence-bond terms as one made with a carbon pure 2p orbital. (A second p orbital would be used to form a K bond to Ta, and the third would be hybridized with s for the u bonds between C, and T a and C, and R.) Both IJcH, and V C H , understandably would be unusually low in such a circumstance. We can therefore postulate that intraligand or interligand steric effects which open up the M-C,-Cp angle in an alkylidene complex will contribute to low values for IJcH, and VCH,,.

A recent neutron-diffraction studyI5 of an electron-deficient neopentylidene complex, [Ta(CHCMe3)C13(PMe3)]2 ( I J c H , = 101 Hz, V C H , = 2605 cm-I), has shown that the Ta-C,-Cp bond angle is large (161’), the M-C,-H, angle is consequently small (85’), and the C-H, bond is stretched to 1.131 (3) A, longer than that predicted (1.120 A) for a pure p-type C-H bond. The electron-deficient T a is believed to be attracting the CH, electron pair (4), or, alternatively, the a hydrogen as a hydride (He-). A more extreme description is a three-center two-electron (or a total of six electrons) bond in 5. It would seem that the hydride would not actually transfer to the metal (in this case) since an additional electron pair is needed to form 6 which the metal does not have. (One could

H

-I (L)Cl,Ta=C-R /‘ 4

H /H\

(L)CI,Ta= C-R

-#+

I

(L)Cl, TaaC-R

5

6

avoid this problem by using descriptions such as (L)C13(H)-

Ta=C+-R or (L)C13(H)Tatr-R.) In essence, Ta and C, are battling for H,-. This electronic driving force operates synergistically with the steric driving force to open up the M-C,-Cp angle. This situation should exist in general whenever IJcH, and V C H , are low (or vice versa). Furthermore, we might expect any increase in the metal’s Lewis acidity to encourage M/H,- interaction. Preliminary X-ray structural datalba suggest that T a C p ( C H C M e X l 2 is a pseudotetrahedral species in which the neopentylidene ligand plane bisects the CI-Ta-CI angle (an orientation the same as the isopropylidene ligand in MnCp(CO)2(CMe2)16b).The Ta-C,-Cp angle is larger still [165 (3)’] and the CMe3 group points up “toward” the C p ring, presumably in order to allow H, (not located) to approach the metal from below (7). Presumably both the steric and

0

l65”(0)

7

electronic driving forces operate in this case as well. Since IJcH, and V C H , are even lower than in [Ta(CHCMe3)C13(PMe3)]2 we can place some confidence in the inverse correlation of each with the M-C,-Cp angle, or, alternatively, the direct correlation of each with the M-C,-H, angle. If we

3217

take the Me3C-C-H angle to be 1 14’ (as in [Ta(CHCMe3)C13(PMe3)]2) then the M-C,-H, angle becomes 81’ in 7. The data in Table IV support the above hypotheses surprisingly well. For example, V C H , in TaCp(CHCMe3)Br2 is 40 cm-] lower than in TaCp(CHCMe3)Clz, and in TaCp”(CHCMe3)Brl it is 15 cm-’ lower than in TaCp”(CHCMe&I2. Ta in the bromide complexes is probably more electrophilic, for the same reason that AIBr3 is a better Lewis acid than AlC1317aand BBr3 is better than BC13, which is better than BF3;17b Br is a poorer ?r-electron donor.17cApparently any steric effect on substituting CI with Br (which would reduce 8 ) is negligible. Note also that I J C H ,is relatively insensitive in this range; it is essentially the same in all four complexes. Replacing C p with Cp” decreases V C H , from 2510 (in C12). This must be due to steric interaction between CMe3 and the methyl groups on Cp” (which increase 8). The electronic effect should be the opposite since Cp” should donate electron density to the metaL9 Apparently the steric effect is more important here. One puzzling result is the effect of replacing a chloride ligand with a neopentyl ligand; 8 should decrease ( I J c H , and V C H , should increase) for both steric and electronic reasons. On replacing one chloride V C H , does increase slightly but IJcH, decreases. On replacing the second chloride, ‘ J c H , remains constant but V C H , decreases slightly. On replacing both chlorides with methyl groups both IJcH, and V C H , decrease markedly. The latter data would suggest that the overall result of replacing C1 with CH3 is to increase 8, Le., to decrease (!) the electron density on Ta. A similar increase in 8 on replacing CI with a neopentyl group would be counteracted by a steric effect to decrease 8. This may be the origin of the small and inconsistent changes in IJcH, and V C H , in the neopentyl derivatives. The effect of replacing Ta with N b is to increase IJcH, and V C H , , consistent with N b being a poorer Lewis acid than Ta.I8 Interestingly, the overall effect on ’ J c H , and V C H , of adding PMe3 to TaCp(CHCMe3)C12is negligible. The concept of the metal interacting with the alkylidene CH, electrons helps explain why H, in an alkylidene complex can be removed so readily, either by another alkyl ligandIga or by an “external base”,lgbto give an alkylidyne complex. One view is that the C-H, bond is long, weak, and susceptible to homolytic cleavage, Le., “R- abstracts H..” An alternative is that a “semibridging” a-hydrogen atom should be more readily removed by a nucleophile (R-), since bridging hydrogen atoms in other electron-deficient compounds (boron hydrides, for example) clearly are.2o W e now go one step further and suggest that intramolecular abstraction of an a-hydrogen atom from an alkyl ligand to give an alkylidene ligand is a related reaction which is facilitated by steric and electronic effects of the type which we have seen above. There is some evidence that this is the case. M-C,-Cp angles in bulky alkyl complexes are known to be larger than expected (e.g., 128’ for neopentyl in Ta(CH2CMe3)3[C(CMe3)(Li-dmp)]19b)and ~JcH, smaller (107 H z in T~(CHZCM~~)~(CHC 105 M ~(axial) ~ ) ; ~ and ~ 112 H z (equatorial) in trigonal bipyramidal Ta(CH2CMe3)4C13a). ‘JcH, is also small ( ‘ J c H = ~ 100 f 2, ~ J c = H 110 ~ f 2 Hz) in neopentyl/neopentylidene derivatives such as TaCp(CHCMe3)(CHzCMe3)2, and for the neopentyl a protons in cis-TaCp(CH2CMe3)(CH2Ph)C12. (No reliable data for V C H , in such compounds are yet available.) There is no way of knowing a priori whether these effects are essentially all steric or a mixture of steric and electronic (but see below for comparison with phosphorus compounds). I n any case we might expect that a methyl ligand would be the least likely (compared to a benzyl or a neopentyl ligand) to give up an

321 8

Journal of the American Chemical Society

a-hydrogen atom since its H, atoms would be furthest from the metal and presumably would interact least. Intramolecular elimination of methane from a methyl complex to give a stable methylene complex does seem to be the most difficult in a series of analogous methyl, benzyl, and neopentyl complexes.6,10 Another reason why this postulate seems reasonable is that a protons in a neopentylphosphonium salt are unexpectedly acidic compared to those in an ethylphosphonium salt.2i However, this must be a slight steric effect which is magnified in phosphonium by the overall positive charge since JCH, salts,22including [Ph3PCH2CMe3]+,23is >125 Hz. Note that I JCH,is larger in alkylidenephosphoranes,22 including Ph3P=CHCMe3.23 Phosphorus(V), unlike Ta(V) or Nb(V), cannot be called electron deficient, and no attraction of the CH, electrons would be expected. Also, the phosphonium salts and alkylidenephosphoranes which are readily available (including Ph3P=CHCMe3) are not crowded compared to most of the Ta and N b species which we have prepared. The JCH, values are therefore “normal”. If we now further postulate that the leaving group must be close to the activated a-hydrogen atom in order to remove it in an essentially concerted process, i.e., that a-abstraction takes place in the cis form of TaCpR2C12 only, then a great many of our findings can be explained. Since cis/trans interconversion is fast relative to the rate of a-abstraction the latter will depend on K,, between the cis and trans forms. K,, can be observed only in the “model” compounds, TaCp(CHzCMe3)(CH?Ph)X2. We must assume that some reasonable amount (say 1%) of cis-TaCp(CH2CMe3)2X2 is present (Keq(t/c) 0 100). A good example of the importance of cis alkyl groups (we propose) is the substantial decrease in rate on replacing Cp by Cp”. We would not expect (according to our comparison of TaCp”(CHCMe3)Cl2 with TaCp(CHCMe3)C12) the Cp” group to differ as markedly from the Cp group electronically as sterically. Since we have found no evidence of a cis form for any TaCp”R2X2 complex we propose that the bulkier Cp” group “locks” the alkyl groups into a relative trans orientation and that very little cis form is ever present. It is interesting to note that intermediate TaCp”(CH2CMe3)3CI (where no “trans” equivalent is possible) must decompose very readily to TaCpt’(CHCMe3)(CH2CMe3)C1. The increase in rate on substituting Br for C1 can also be accounted for in part by a shift in K,, since it varies in the opposite sense in TaCp(CHzCMe3)(CH2Ph)X2 when X = Br (more cis at higher T ) than when X = CI (more trans at higher T). Even some of the larger K B ~ / K cratios I (-500) could be explained in this manner. However, we are convinced that there must be a significant electronic difference between Br and CI which parallels that found in the neopentylidene complexes; an a-hydrogen atom in cis-TaCp(CHzCMc3)2Br2 should be more “activated” than that in ci~-TaCp(CH2CMe3)2C12since the metal should be more electrophilic. This effect could be fairly significant and still be masked by the cis/trans effects. In a case where no isomers are possible (Ta(CH2CMe3)dX) we do know that the complex is many times more stable when X = OCMe3 (a better r-electron donor than Cl) than when x = c1.3a Finally, the solvent effect also is at least qualitatively consistent with a shift in the cis/trans equilibrium toward the more polar cis form in more polar solvents like chloroform. The relative amounts of cis forms in two solvents can easily be 20 or 30. However, in the case where little or no trans form is present (TaCp(CH2CMe3)(CH2Ph)Br2) the solvent effect (chloroform vs. ether) is still significant ( = 5 ) , It is interesting to note that little solvent effect (chloroform vs. benzene) was observed in another instance where no isomers are pos~ible.~~ We prefer to view the abstraction reaction as a concerted



1 101:12

June 6, 1979

process. However, if we had to describe M-CH2R cleavage as a separable step it might be best described as a homolytic cleavage for two reasons. First, light dramatically accelerates the a-abstraction reaction. (Light is believed to most characteristically induce homolytic M-R bond cleavage.) Dineopentyl could form from neopentyl radicals generated from the trans isomer since no suitable activated a hydrogen is present. Second, metal reduction is a logical alternative if no suitably activated H, is close by. (NbCp(CH2CMe3)2C12 may be reduced to Nb(1V) since, as we have seen, H, is not as activated in N b C H C M e 3 complexes as in Ta=CHCMe3 complexes.) Of course the M-CH2R bond strength will depend on what other ligands are present, how the CH2R group is oriented relative to them, and what R is. Interestingly, Lappert found that a M-CH2CMe3 bond was surprisingly weaker than either a M-CH2Ph or M-CH2SiMe3 bond in M(1V) alkyls such as Zr(CH2R)4.25We might also add that one might expect M-alkyl bond strengths in crowded molecules to be weaker, all else being equal. An intriguing possible partial explanation of the substantial isotope effect is that H prefers to be in the “activated” (possibly “semibridging”) position between C, and Ta much more so than D owing to the greater zero-point energy of the terminal C-H bond (a “thermodynamic” isotope effect).27 A small energy difference between the two (AG = 0.4 kcal mol-’) would give K,, = 2. The true kinetic isotope effect in the a-abstraction step would then only have to be on the order of 3. There is no way of separating the two possible types of isotope effects using the data we have.28 We can now understand the importanccof structure in dialkyl cyclopentadienyl complexes. TaCp(CHzCMe3)(CH2Ph)X2 can attain the cis form but it is in that form not crowded enough for rapid a-abstraction. TaCp(CH2CMe3)2X2 can still attain the cis form and is crowded enough for a-abstraction. TaCp”(CHzCMe3)2Xz is so crowded that very little cis form is present, and a-abstraction is therefore relatively slow. When TaCp(CH2CMe3)(CHzPh)Br2 decomposes it gives primarily toluene and TaCp(CHCMe3)Br2, exactly the opposite of what one would predict based on the presence of the phenyl ring. A neopentyl LY hydrogen is more abstractable. For steric reasons the M-C,-C@ angle in the benzyl ligand will not be as large. In fact M-C,-C@ angles in known electron-deficient complexes such as Zr(CH2Ph)4 are abnormally small (as low as 90°) compared to those in Sn(CH2Ph)4, for example.29 Interaction of the metal with the ortho C-H bond in the phenyl ring, or with the ring itself (another manifestation of a metal’s electron deficiency), was the explanation. This type of interaction could actually be partly or even largely responsible for the apparent relatively slow a-abstraction from a benzyl a-carbon atom under most circumstances; the metal acquires electron density by interacting with the phenyl ring instead of with the a-hydrogen atom.30 The obvious question is whether some version of an a-abstraction reaction is possible for other, and in particular later, transition metals. Certainly electron-deficient metals will interact with “hydride” bound to carbon, in Mo pyrazolylborate complexes,31 in protonated diene-Fe(0) c o m p l e ~ e s , and ~~,~~ in O S ~ ( C O ) I O ( C H ~ ) ( H[One ) . ~ ~interesting piece of data should be compared with that which we have presented here; in [Fe(CO),(C4H,)]+ JCHin the C-“bridging H”-Fe32a system is 74 H z . ~ Therefore ~ ~ ] an electron-deficient metal could in theory set up H,- in an alkyl complex to be abstracted or could itself accept H, to give an alkylidene/hydride complex ( “ a - e l i m i n a t i ~ n ” ~from ~ ) which the hydride is subsequently removed; the difference between the two processes may be more formal than real. If this is true then it is peculiar that, although what is overall an a-abstraction is also known for Mo and W neopentyl complexes,’ neopentyl ligands bound to R u , ~ ~

Wood, McLain, Schrock

/

Monocyclopentadienyl Mononeopentylidene Complexes

Rh,35 or Pt36 lose a y-hydrogen atom to give P,P-dimethylmetallacyclobutane complexes. Experimental Section All manipulations were done under N2, either in a Vacuum Atmospheres HE43-2 drybox or by standard Schlenk techniques. Ether and toluene were distilled from sodium benzophenone ketyl under nitrogen. Pentane was purified by an acid wash and then distilled from LiAIH4 under nitrogen. Acetonitrile and benzonitrile were distilled from P205 (the latter at reduced pressure) after refluxing for 2 days. Metal halides and thallium reagents were purchased from standard sources and sublimed prior to use. LiCH2CMe3,3a M g ( C H 2 C M e 3 ) 2 ( d i o ~ a n e ) ,Ta(CH2CMe3)2C13,3a ~~ Nb(CH2CMe&C13,6 z n ( C H ~ P h ) 2 , and 3 ~ LiC5Mes3*were prepared by published methods. Elemental analyses were done by Alfred P. Bernhardt or by Schwarzkopf, Chloride analyses were usually done i n these laboratories by titration of digested samples with AgN0339 or Hg(N03)2.40 ' H N M R spectra were done at 60 and 90 (PerkinElmer) or 270 MHz (Brucker) and I3C spectra at 15 (JEOL) or 67.89 MHz (Brucker). 1. Preparation of Zn(CDzCMe3)zand Zn(CDHCMe3)z.A solution of LiCD2CMe3 (4.0 g) in 30 mL of ether was added slowly to 5.78 g of ZnC12(dioxane) in 50 mL of ether with stirring. After 2 days, the mixture was filtered and the ether was removed in vacuo. The residue was extracted into pentane and the solution filtered. The pentane and residual dioxane were removed in vacuo (5 h). The crude yellow liquid u a s distilled at 76 "C (20 mm) to give 3.16 g (58%) of colorless Zn(CD2CMe3)2.This is an improvement of the previously published procedure.'" Zn(CDHCMe3)2 was prepared similarly. 2. Preparation of Ta(CHZCMe3)Xd. A solution of 3.1 1 g of Zn(CH>CMe3)2in 25 mL of toluene was added slowly to 16.12 g of TaC15 (50% excess) i n 75 mL of toluene with stirring. The reaction mixture was stirred for 1 week at room temperature. The mixture was filtered and the toluene removed in vacuo. The residue was extracted into pentane and the solution filtered. Removing the pentane in vacuo yielded 10.8 g of yellow, crystalline Ta(CH2CMe3)CI4 (92%). I t can be recrystallized from pentane ( 2 6 0 % recovery). It decomposes at room temperature in several days in absence of TaC15. Anal. Calcd for TaC5Hl IC14:CI, 36.00. Found: Cl, 35.40. IH NMR ( 7 , C6D6): 6.68 (s, 2, CH2CMe3), 9.04 (s, 9, CH2CMe3). Ta(CH2CMe3)Brd (7 7.04,9.00) is prepared analogously ( 1 day at 60 "C). 3. Preparation of Ta(CHzCMe3)2Br3.Zn(CH2CMe3)2 ( 1 .O g) in I5 mL of toluene was added slowly to a vigorously stirred suspension of 4.0 g of TaBrs (excess) in 50 mL of toluene over a period of 1 h. The mixture was filtered and all toluene removed in vacuo leaving a greenish-orange oil. This was dissolved in pentane and the solution was treated with a small amount of activated charcoal and filtered. The yellow solid which remained after removing all pentane from the filtrate in vacuo was pure Ta(CH2CMe3)2Br3 according to its ' H N MR spectrum in C6H6 ( 7 6.90 (s), 8.60 (s), 2:9 ratio) compared to mixtures containing Ta(CHrCMe3)lBrz (7 7.20 (s), 8.50 (s)) and/or Ta(CH2CMe3)Br4 (7 6.80 (s), 8.70 (s)); the shiftsvary somewhat with absolute and relative concentrations of the three. Usually the product is a partially crystalline, yellow-orange oil containing Ta(CHzCMe3)3Brz as the main contaminant (it is comparatively unreactive) since the result sensitively depends on the purity of the TaBrs, how finely divided it is, the rate of Zn(CH2CMe3)2addition, etc. I n that case thepurity was determined by IH N M R and stoichiometries for subsequent reactions were adjusted accordingly. Little Ta(CHzCMe3)2Br3 is obtained in pentane since TaBr5 is virtually insoluble in this solvent. 4. Preparation of Ta(CH2CMe3)(CH2Ph)X3.A solution of 1.26 g of Zn(CHzPh)2 in 20 mL of toluene was added slowly to 4.0 g of Ta(CH2CMe3)C14 in 30 mL of toluene with stirring. The mixture was stirred for 1 h and the solution filtered. The toluene was concentrated to 5 mL in vacuo and 20 mL of pentane added. Storing the solution at -20 "C for 2 days yielded 3.50 g of red-orange, crystalline Ta(CH2CMe3)(CHzPh)CI3 (80%). Anal. Calcd for T ~ C I ~ H I C, ~ C32.06; I ~ : H, 4.03; CI, 23.66. Found: C, 31.69; H, 3.73; CI, 23.58. ' H N M R ( 7 , C6D6): 2.53-2.96 (m, 5, C6H5), 6.36 (s, 2, CH2Ph). 7.59 (s, 2, CH2CMe3), 8.93 (s, 9, CH2CMe3). Ta(CHzCMe3)(CHzPh)Br3 was prepared analogously (7 6.36, 7.40, 8.88).

3219

5. Preparation of TaCp(CH2CMe3)CI3. Ta(CHzCMe3)CI4 (3.25 g) and TlCp (2.22 g) were stirred in 50 mL of toluene for 12 h. The mixture was filtered and the toluene concentrated to I O mL. Pentane was added and 3.10 g of orange, crystalline TaCp(CHzCMe3)Cl3 isolated by filtration (89%). 'H N M R ( 7 , C6D6): 4.34 (S, 5, c p ) , 8.12 ( s , 2, CH2), 8.63 (s, 9, CMe3). 6. Preparation of TaCpt'(CHzCMe3)CI3.Ta(CH2CMe3)C14 (7.92 g) and LiC5Mes (2.86 g) were stirred in 50 mL of ether for 2 h. The mixture was filtered and the ether removed from the filtrate in vacuo. The residue was extracted into toluene and the solution filtered. The toluene was removed in vacuo and the crude product recrystallized from ether to give 5.54 g of red-orange crystals. The ether was removed from the filtrate in vacuo and the residue was again extracted into toluene. The solution was filtered and the toluene concentrated until crystals were noted. An equal volume of pentane was added and the solution was stored at -20 "C for 24 h to give another 2.0 g of compound (overall yield 76%). IH N M R ( 7 , C6D6): 8.03 (s, 15, CsMes), 8.54 (s, 9, CH2CMe3), 8.62 (s, 2, CH2CMe3). 7. Preparation of TaCp(CH2CMe3)2CIzand TaCp(CDzCMe3)zCIz. A solution of 1.41 g of Ta(CH2CMe3)2C13 in 30 mL of ether at 0 "C was treated with 0.88 g of TICp. The reaction mixture was stirred for 30 min and filtered. Removing the ether in vacuo yielded 1.2 g of dark-red, crystalline TaCp(CH2CMe3)2C12(80%). Recrystallization from pentane yielded 0.80 g (53%). TaCp(CD2CMe3)2C12 was prepared similarly, mol wt (cyclohexane) 433 (calcd, 463). Anal. Calcd for TaC1sH27C12: CI, 15.44. Found: CI, 15.51. IH N M R ( 7 , C6D6): 4.24 (s, 5, Cp), 8.22 ( s , 4, CH2), 8.74 (s, 18, CMe3). TaCp(CH2CMe3)2C12 was also prepared by reacting 1.0 g of TaCp(CHfCMq)CI3 with 0.30 g of Mg(CH2CMe3)2(dioxane) in 30 mL of ether. After 20 min the solution was filtered and the ether removed in vacuo. The compound was recrystallized from pentane (0.61 g, 56%).I3C N M R (ppm, toluene-d8, IH gated decoupled): 117 (t, CH2CMe3, ' J C H = 118 Hz), 116 (d, Cp, ' J C H= 175 Hz), 37.2 (5, CH2CMe3), 35.1 (q,CH2CMe3, ~ J C H = 124 Hz). 8. Preparation of TaCpf'(CH2CMe3)2CI2.Ta(CH2CMe3)2CI, (2.00 g) and LiCSMe5 (0.66 g) were stirred in 30 mL of ether for 2 h. The mixture was filtered and the ether removed in vacuo. The residue was extracted into toluene and the mixture filtered. The toluene was removed in vacuo and the crude product recrystallized from ether to give deep red crystals (1.73 g, 70% yield). Anal. Calcd for TaC2oH37C12: C, 45.38; H , 7.05; CI, 13.40. Found: c , 44.90; H , 7.13iC1, 13.52. ' H N M R ( T , C ~ D ~8.23 ) : (S, 15, CSMes), 8.77 (s, 18, CH2CMe3), 9.09 (s, 4, CH2CMe3). l3C N M R (ppm, C6D6, IH gated decoupled): 124 (s, CSMeS), 117 (t, CHzCMe3, ~ J C H = 116 Hz),37.7 (s,CH2CMe3),36.1 (q,CH3CMe3,lJCH = 123 Hz), 12.8 (q, C5Me5, ~ J C H = 128 Hz). 9. Preparation of TaCp"(CH2CMe3)zBrz.Pure Ta(CHzCMe3)2Br3 (2.70 g) was dissolved in 50 mL of ether and 0.70 g of LiCsMes added. The flask was covered with foil and stirred for 1 h. The ether was removed in vacuo and the residue was extracted into 125 mL of pentane. The extract was filtered and stood at -30 "C overnight to give 1.45 g of small, red needles. Decreasing the volume afforded 0.35 g of additional product, total yield 1.80 g (56%). The compound in benzene is red by transmitted light but blue-purple by reflected light. ' H N M R ( 7 , C6D6): 7.96 (S, 15, C5Me5), 8.54 (s, 18, CH2CMe3). 9.35 (s,4, CH2CMe3). 10. Preparation of TaCp(CHzCMe3)(CHzPh)X2.Ta(CH2CMe3)(CHzPh)C13 (3.3 g) and TlCp (1.98 g) were stirred in 50 mL of toluene for 30 min. The mixture was filtered and the toluene removed in vacuo. Pentane was added and 2.62 g of deep-red, crystalline TaCp(CH2CMe3)(CH2Ph)CI2isolated by filtration (75%). Anal. Calcd for TaC17H23C12:CI, 14.80. Found: CI, 14.80. IH N M R ( 7 , CDC13, -50 "C): cis isomer 2.68-3.18 (m, 5, C6H5), 3.57 ( S , 5, c p ) , 6.80 (d, 1, CHAHgPh, 'JHH = 12 Hz), 7.24 (d, 1, , = 14 CHAHBPh, ' J H H = 1 2 Hz), 7.74 (d, 1, c H ~ H e c M e 3 'JHH HZ), 8.37 (d, 1, C H A H B C M ~'J" ~, = 14 Hz), 8.82 (s, 9, CH2CMe,), trans isomer 2.68-3.18 (m, 5, C6H5),3.46 (s, 5, Cp), 7.35 (s, 2, CHlPh), 8.50 (s, 2, CH2CMe3), 9.13 (s, 9, CHZCMe3). I3C N M R (ppm, CDC13, ' H gated decoupled, -50 "C): cis isomer 149.8 (s, Clps0),127.8, 126.8, 123.9 (d, other phenyl carbons, ~ J C H = 159 Hz), 116.4 (d, Cp, IJCH 181 Hz), 114.0 (t. CH2CMe3, ~ J C = H 112 Hz), 96.0 (t, CH2Ph, ~ J C H= 126 Hz), 37.1 (s, CH2CMe3), 34.2 (q,

3220

Journal o f t h e American Chemical Society

CHzCMe3, I J C H = 124 Hz), trans isomer 143.0 (s, Cipso), 130.5, 126.7, 123.7 (d, other phenyl carbons, ~ J C = H 159 Hz), 1 15.6 (d, Cp, ' J C H = 181 Hz), 114.2 (t, CH2CMe3, ' J C H % 112 Hz), 98.6 (t, CH2Ph, ~ J C=H128 Hz), 35.1 (s, CH2CMe3), 34.0 (q, CHzCMe3, 'Jclj = 124 Hz). TaCp(CHzCMe3)(CHzPh)Brzwas cleanly prepared in situ in ether or CDC13 at -20 "C for determination of its rate of decomposition at 32 OC. In ether k = 0.029 min-l (rip = 24 min). I n CDC13 k = 0.155 min-' ( [ I / > = 4.5 min). Virtually the only decomposition products were TaCp(CHCMe3)Brz and toluene. ' H N M R (7,CDC13, -50 "C): cis isomer 2.54-3.34 (m, 5, CsHs), 3.53 ( S . 5, c p ) , 6.83 (d, 1, CHAHBPh, ' J H H = 1 1 Hz), 7.26 (d, I , CHAHBPh, 'JHH = 1 I Hz), 7.83 (d, I , C H A H B C M ~'J"~ . = 15 Hz), 8.59 (d, 1, CHAHBCMe3, 'J" = 15 Hz), 8.80 ( s , 9, CHzCMe3). 1 1 . Preparation of TaCp"(CH2CMe3)(CH2Ph)Clz. Ta(CH2CMe3)(CH2Ph)C13 (4.85 g) and LiCsMes (1.53 g) were stirred in 50 mL of ether for 2 h. The mixture was filtered and the ether removed in vacuo. The residue was extracted into toluene and filtered. The toluene was removed in vacuo and the crude product recrystallized from a pentane/ether (25/75) solution. Deep-red crystals (4.6 g) were isolated (78%). Anal. Calcd for TaC22H33CI2: CI, 12.91. Found: CI, 12.88. IH N M R ( 7 , ChDh): 2.42-3.29 (m, 5, phenyl), 7.80 (s, 2, CH2Ph), 8.13 (s, 15, C5Me5). 8.83 (s, 2, CHzCMe3), 8.86 (s, 9, CHzCMe3). 12. Preparation of TaCp(CHCMe3)Clz and TaCp(CDCMe3)CIz. Ta(CH2CMe3)2C13 (4.14 g) and TlCp (2.60 g) were stirred in 30 mL of toluene for 24 h. The solution was filtered and the toluene removed in vacuo. The residue was extracted into 25 mL of a pentane/toluene (80/20) solution. Storing the solution at -20 OC for 24 h yielded 2.85 g of red, crystalline TaCp(CHCMe3)C12 (76%). Anal. Calcd for TaCloHl~C12:C, 31.03; H, 3.91; CI, 18.32. Found: c, 31.17; H, 4.05; CI, 18.30. ' H N M R ( 7 , C6D6): 3.62 (s, I , CHCMe3). 3.45 (s, 5, Cp), 9.05 (s, 9, CHCMe3). I3C N M R (ppm downfield from Me& C6D6, ' H gated decoupled): 246 (d, CHCMe3, ~JCH = 84 Hz), 107 (d, Cp, ~ J C H = I80 Hz), 48 (s, CHCMej), 33 (q, CHCMe3, ~ J C H= 126 Hz). Mol wt (cyclohexane): 404 (calcd, 388). A side product was isolated by extracting the TIC1 residue with acetonitrile. Black crystals were obtained by cooling the deep-green acetonitrile solution to -30 "C for 12 h. This compound was identified as TaCp2Clz by analysis and its infrared spectrum.41Anal. Calcd for TaCloH1oCl2: C, 31.44; H, 2.64; CI, 18.57. Found: C, 31.69; H, 2.52; CI, 18.47. TaCp(CDCMe3)Clz was prepared as follows. A solution of 0.25 g of TaCp(CDzCMe3)zCIz in 5 mL of toluene was stirred for 4 h at 60 "C. The toluene was removed in vacuo and the residue extracted into 5 mL of a pentane/toluene (80/20) solution. Storing the solution a t -20 OC for I day yielded 0.1 1 g of red, crystalline TaCp(CDCMe3)Cl2 (52%). 13. Preparation of TaCp(CHCMe3)Brz and Observation of TaCp(CHzCMe3)zBrt.A mixture of 72% Ta(CHzCMe3)2Br3 and 28% Ta(CHzCMe3)3Br2 (2.1 g; 1.52 g of Ta(CH2CMe3)2Br3 = 2.7 mmol) was dissolved in toluene and treated with 0.73 g (2.7 mmol) of TlCp at, 25 "C with rapid stirring. The solution turned purple in I O s, then red within 1 min. After I h TIC1 was filtered off and all toluene removed in vacuo. The solid was triturated with pentane (5 mL) and 0.72 g of TaCp(CHCMe3)Brz was filtered off. Standing the filtrate at -30 "C gave an additional 0.10 g (64% total yield based on Ta (CH 2C Me3)2 Br 3). TaCp(CH C Me3) Br2 can be recrystallized from I :4 toluene/pentane mixtures at -30 "C (cf. TaCp(CHCMe3)Cl2 above). Anal. Calcd for TaCloHlsBr2: C , 25.23; H, 3.17. Found: C, 25.51; H. 3.50. ' H N M R ( 7 , C6D6): 3.56 (s, 1 , CHCMe3), 4.35 (s, 5, Cp), 9.03 (s, 9, CHCMe3). I3cN M R (ppm, C6D6, gated ' H decoupled); 254 (d, CHCMe3, ' J C H = 83 Hz), 106 (d, Cp, ' J C H = 180 Hz), 48.3 (s,CHCMe3), 32.5 (q,CHCMe3, I J C H = 126 Hz). Intermediate TaCp(CHzCMe3)2Br2 could be observed by ' H N MR in the aliquot of a reaction done in ether at ca. -25 OC for 15 min. The Cp signal was located 3.50 ppm downfield of the low-field ether quartet resonance; the Cp resonance in TaCp(CHCMe3)Brz is found at 3.27 ppm downfield of the quartet. The rate of decornposition of the former to the latter was obtained at 2 and -1 1 "C by estimating the half-life of the former at each temperature (7 f 1 and 32 f 2 min, respectively). Reaction of Ta(CHDCMe&Br3 with TlCp in toluene gave a

/

IOI:12

/

June 6 , 1979

mixture of 85% TaCp(CDCMe3)Brz and 15% TaCp(CHCMe3)Brz (by IH N M R ) or k H / k D = 6 f I . 14. Preparation of TaCp"(CHCMe3)Clz. Four solid samples of TaCp"(CH2CMe3)(CH2Ph)Clz (0.25 g each) were heated at 145 "C for 20 min and then sublimed at 0.05 p for 1 h. Each sublimate was dissolved in 2-3 mL of pentane. These were combined and stored at -20 "C for 2 days. The solution was filtered and the pentane removed in vacuo to give a red oil (0.18 g, 23% yield, >90% pure). ' H N M R ( 7 , C6D6): 5.13 ( S , 1, CHCMe3), 8.01 ( S , 15, CsMes), 8.84 (s, 9, CHCMe3). I3C NMR (ppm downfield of Me&, C6D6, IH gated decoupled): 242 (d, CHCMe3, ' J C H= 83 Hz), 119 (s, CSMes), 47.1 ( s , C H C M q ) , 33.7 (q, CHCMe3, ~ J C H= 127 Hz), 12.3 (q, CSMes, ' J C H = 128 Hz). 15.Preparation of TaCp"(CHCMe3)Br2.T ~ C P ' ' ( C H ~ C M ~ ~ ) ~ B ~ ~ (0.80 g) was dissolved in minimal CDC13 and the solution was stood in the dark for 2 days. The solvent was removed and the residue was doubly recrystallized from minimal pentane at -30 "C, yield 0.40 g (57%). 'H N M R (7, C6D6): 5.18 (S, 1 , CHCMe3), 7.88 ( S , 15, CSMes), 8.83 (s, 9, CHCMe3). I3cN M R (ppm, C6D6, ' H gated decoupled): 249 (d, CHCMe3, ' J C H = 77 Hz), 119 (s, CsMes), 48.2 (s, C H C M ~ ~ ) , ~ ~ . ~ ( ~ , C H= C126 M Hz),13.2(q,CsMes, ~~,'JCH 'JCH = 128 Hz). 16. Preparation of Nb(CsHdMe)(CHCMe3)CI2 and NbCp(CHCMe3)CIz. Nb(CH2CMe3)2C13 (0.76 g) and TICsH4Me (0.63 g) were stirred in 20 mL of toluene for I O min. The mixture was filtered and the toluene removed in vacuo. The residue was extracted into pentane and the solution filtered. Storing the filtrate at -20 "C for 24 h yielded 0.23 g of purple, crystalline NbCp'(CHCMe3)C12 (30%). ' H N M R ( 7 , C6D6): 2.70 (s, I , C H C M q ) , 4.02 (m, 2, CsH2H'2Me), 4.42 (m, 2, CsHzHz'Me), 8.20 (s, 3, CsHdMe), 9.03 (s, 9, CHCMe3). I3C N M R (ppm downfield of Me4Si, C6D6, IH gated decoupled): 254 (d, CHCMe3, ~ J C H= 95 Hz), 109 (d, CzH2Cz'H2CMe, ~ J C H = 177 Hz), 105 (d, CzHzC2'H2CMe, ' J C H = 180 Hz), 48.7 (s, CHCMe3), 30.7 (q, CHCMe3, ~ J C H = 125 Hz), 14.9 (4. CsHdMe, ' J C H = 128 Hz). NbCp(CHCMe3)Cl2 was prepared similarly. TlCp reacted slowly compared to TICsH4Me. Reaction times ranged from 4 to 12 h and yields from 7 to 15%. Anal. Calcd for NbCloHlsC12: C, 40.17; H, 5.05; CI, 23.71. Found: C, 40.19; H , 5.16; C1, 23.66. ' H N M R ( 7 , C&): 2.75 ( s , I , CHCMe3), 4.20 (s, 5, Cp), 9.05 (s, 9, CHCMe3). 17. Preparation of TaCp(CHCMe3XCHzCMe3)CI. A solution of LiCH2CMe3 (0.20 g) in I O mL of pentane was added slowly to 1 .O g of TaCp(CHCMe&l2 in 30 mL of pentane with stirring. The mixture was filtered and the pentane removed in vacuo to yield 0.72 g of a red-orange oil (66%). ' H N M R ( 7 , C6D6. 270 MHz): 4.08 ( s , I , CHCMe3), 4.27 ( s , 5, c p ) , 7.91 (d, 1, C H A H B C M ~J ~=, 13 Hz), 8.90 (s, 9. CHCMe3), 8.95 (s, 9, CH2CMe3), 9.28 (d, 1, C H A H B C M C ~J , = 13 Hz). 13c N M R (ppm downfield from Me4Si, C6D6, ' H gated decoupled): 240 (d, CHCMe3, ' J C H= 81 Hz), 104 (d, Cp, ~ J C = H 177 Hz), 79.1 (dd, = 100, 1 I O f 2 Hz), 47.3 (s, CHCMe3). 34.8 C H A H B C M ~~ ~J ,C H (s, CH2CMe31, 34.6 (q, CHzCMe3, ' J C H = 125 Hz), 33.4 (q, CHCMe3, ' J C H = 125 Hz). 18. Preparation of TaCp"(CHCMe3)(CHzCMe3)CI. A solution of 2.68 g of LiCH2CMe3 was added slowly to 8.50 g of TaCp"(CH2CMe3)CI3 at -78 "C with stirring. The mixture was warmed slowly to room temperature and stirred for 2 h. It was filtered and the pentane was removed from the filtrate in vacuo. The crude compound was recrystallized from pentane and 6.14 g of orange, crystalline TaCp"(CHCMe3)(CH2CMe3)CI was isolated (7 1%). Anal. Calcd for TaC2,#&1: C , 48.73; H, 7.36; CI, 7.19. Found: C, 48.25; H, 7.51; C1, 7.40. ' H N M R ( 7 , C6D6. 270 MHz): 6.27 (s, I , CHCMQ), 8.1 2 (s, 15, CSMes), 8.65 (s, 9, CHCMe3), 8.75 (s, 9, CH2CMe3), 9.08 (d, 1, C H A H B C M ~J ~=, 14 Hz), 9.25 (d, I , c H ~ f f ~ C M eJ 3=, 14 Hz). 13CN M R (ppm downfield from Me4Si, H 76 Hz), 116 C6D6, IH gated decoupled): 232 (d, CHCMe3, ~ J C = (s, C5Me5), 80.8 (dd, C H A H B C M ~ ~J ,C H = 100, 110 f 2 Hz), 47.3 (s, CHCMe3), 35.2 (q, CH2CMe3, IJCH = 123 Hz), 34.1 (q, = 124 Hz), 12.3 (q, CsMes, ' J C H= 127 Hz). CHCMe3, ~ J C H 19. Preparation of TaCp(CHCMe3XCHzCMe3)z.A solution of 0.7 1 g of LiCH2CMe3 in 30 mL of pentane was added slowly to TaCp(CHCMe,)CI2 (1.76 g) in 30 mL of pentane with stirring. The mixture was filtered and the pentane removed in vacuo to give 1.75

Wood, McLain, Schrock / Monocyclopentadienyl Mononeopentylidene Complexes g of a yellow-orange oil (84%). ' H N M R (T, C6D6, 270 MHz): 4.39 (s, 5, Cp), 5.01 (s, I , C H C M q ) , 8.44 (d, 2, C H A H B C M ~J~ = , 12 Hz). 8.79 (s, 9, CHCMe3), 8.91 (s, 18, CH2CMe3), 10.52 (d, 2, C H A H B C M ~J ~=, 12 Hz). 13C N M R (ppm downfield from MeoSi, C6D6, ' H gated decoupled): 234 (d, CHCMe3, ~ J C H= 80 Hz), 104 (d, Cp, ~ J C H = 175 Hz), 85.4 (dd, C H A H B C M ~' J~C,H = 100 f 2, 110 f 2 Hz), 47.1 (S, CHCMe3), 35.1 (q, CH2CMe3, ~ J C H = 125 Hz), 34.8 (s, CH2CMe3), 34.0 (q, CHCMe3, ~ J C H= 125 Hz). 20. Preparation of TaCp"(CHCMe3XCHzCMe3)z.A solution of 0.08 g of LiCH2CMe3 in 5 mL of pentane was added slowly to 0.50 g of TaCp"(CHCMe3)(CH2CMe3)CI in 20 mL of pentane with stirring. The mixture was filtered and the pentane was removed in vacuo. The crude product was recrystallized from pentane and 0.32 g of yelloworange, crystalline TaCp"(CHCMe3)(CH2CMe3)2 was isolated (60%). The compound sublimed readily at 100 "C and 0.05 p. Anal. Calcd for TaC2jH47: C, 56.80; H, 8.96. Found: C, 56.33; H, 8.91. IH N M R (T,C6D6, 270 MHz): 6.87 (s, I , CHCMe3), 8.15 (s, 15, CsMej), 8.63 (s, 9, CHCMe3), 8.73 (s, 18, CH2CMe3), 8.86 (d, 2, C H A H R C M ~J ~=, 12 Hz), 10.68 (d, 2, C H A H B C M ~J ~=, 12 Hz). I3C NMR (ppm downfield from Me4Si, C6D6, IH gated decoupled): 228 (d, CHCMe3, ' J C H = 76 Hz), 114 (s, CjMej), 89.5 (dd, C " A H B C M ~~~J, C H = 100, 1 I O f 2 Hz), 47.7 (s, CHCMe3), 35.4 (q, CH2CMe3, ~ J C H= 124 Hz), 34.6 (5, CH2CMe3), 34.1 (q, CHCMe3, ' J C H = 126 Hz), 12.1 (q, CsMej, IJCH = 127 Hz). 21. Preparation of TaCp(CHCMe3)MeCIand TaCp(CHCMe3)Mez. A solution of commercially available methyllithium in ether was treated with 1 equiv of dioxane to precipitate the dissolved LiBr. The "halide-free" methyllithium was titrated with propanol using 1 , I O phenanthroline as an indicator.42 TaCp(CHCMe3)C12 (1.94 g ) was dissolved in 30 mL of T H F and cooled to -78 "C. A solution of halide-free LiCH3 ( 1 equiv per Ta) in 20 mL of T H F was added dropwise over a period of 30 min. At the end of the addition the solution was allowed to warm slowly io room temperature. The T H F was removed in vacuo and the residue was extracted into pentane and filtered. The pentane was removed in vacuo and the resulting orange solid was sublimed at 50 "C and 0.5 p onto a -78 "C probe to give 0.91 g (50%) ofdark-orangecrystals (290% pure). ' H N M R (T,C6D6): 4.12 (s, I , CHCMe3), 4.48 (s, 5, Cp), 8.87 (s,9, CHCMe3),9.47 (s, 3,TaMe). TaCp(CHCMe3)Me2 was prepared similarly using 2 equiv of LiCH3 per Ta. The yield was highly variable (7-69%) and no attempt was made to optimize the procedwe. IH N M R (7,C6D6):4.40 (s, 5, Cp),4.99 (s, I,CHCMe3),8.82 (s,9,CHCMe3),9.88 (s,6,TaMe). I3C NMR (ppm downfield of MedSi, C6D6, IH gated decoupled): 231 (d, CHCMe3, 'JCH = 78 Hz), 104 (d, Cp, ' J C H= 176 Hz), 46.6 (s, CHCMe3), 37.8 (q,TaMe, ' J C H = 118 Hz),33.3 (q,CHCMe3,lJcH = 127 Hz). 22. Preparation of TaCp(CHCMe3)Clz(PMe3).This yellow adduct crystallizes out of benzene or toluene (in which it is slightly soluble) on adding PMe3 to TaCp(CHCMe3)C12. It is extremely soluble in dichloromethane or chloroform. IH N M R (7,CD2C12): 4.01 (d, 1, J H P = 1.6 Hz, CHCMe3), 4.15 (d, 5, J H P= 2.9 Hz, Cp), 8.40 (d, 9, J H P = 8.6 Hz, PMe3), 8.87 (s, 9, CHCMe3). I3C NMR (ppm, CDCI3, = 84, 2 J ~ = p 16 Hz, ' H gated decoupled): 272 (d, CHCMe3, ~ J C H ' H decoupled), 101 (d multiplet, Cp, 'JCH = 179 Hz), 49.2 (s, CHCMej), 31 .O (q, CHCMe3, ' J C H= 126, 4 J ~x p3 Hz, ' H decoupled), 17.0 (dq, PMe3, I J C H = 129, lJcp = 29 Hz). 23. Preparation of TaCp[NC(Me)CHCMe3]CIz. TaCp(CHCMe3)CIz ( 1 .O g) was stirred in 5 mL of acetonitrile for 1 day. The acetonitrile was removed in vacuo and the residue was extracted into minimal ether. Storing the solution at -20 "C for 1 day yielded 0.73 g of orange, crystalline TaCp[NC(Me)CHCMe3]CI2 (66% yield; ratio of E/Z = 1.6). Anal. Calcd for TaC12HlsNC12: C, 33.67; H, 4.24; N, 3.27; CI, 16.56. Found: C, 33.50; H, 4.33; N , 3.21; CI, 16.61. ' H N M R (T, C6D6): E isomer 4.07 (s, 5, c p ) , 5.64 (br s, 1, CHCMe3),8.32 (d, 3, NC(Me), J = 2 Hz), 8.64 ( s , 9, CHCMe3), Z isomer 4.05 ( s , 5, Cp), 4.93 (br s, I , C H C M q ) , 8.28 (d, 3, NC(Me), J = 2 Hz), 8.94 (s, 9, CHCMe3). 13CN M R (ppm, C6D6, IH gated decoupled): E isomer 147 (s, NC(Me)), 130 (d, CHCMe3, J C H = 152 Hz), 112 (d, Cp, ' J C H = 179 Hz), 31.7 (4. CHCMe3, ~ J C H = 125 Hz), 31.7 (s, C H C M e d , 26.5 (q, NC(Me), ~ J C H= 128 Hz), Z isomer 149 ( s , NC(Me)), 134 (d, CHCMe3, ~ J C H = 152 Hz), 112 (d, Cp, ~ J C H= I ~ ~ H Z ) , ~ I . ~ ( ~ , C H C125H~),31.7(s,CHCMe3), M~~,'JCH= 19.5 (4. NC(Me), ' J C H = 128 Hz).

3221

24. Preparation of TaCp[NC(Ph)CHCMe3]Clz.A solution of 0.32 g of TaCp(CHCMe3)Clz and 0.10 g of benzonitrile in toluene was left at 25 "C for 2 days. The toluene was removed in vacuo and the residue was extracted into pentane. Cooling the solution at -20 "C for 1 day yielded 0.25 g of orange and yellow crystals (61% yield; E/Z = I / 3.4). ' H N M R ( T , C6D.5): 2 isomer (orange crystals) 2.49-2.98 (m, 5, Ph), 4.38 (s, 5, Cp), 5.18 (s, I , CHCMe3), 8.57 ( s , 9, CHCMe3), E isomer (yellow crystals) 2.88 (m. 5, Ph), 4.41 (s, 5, Cp), 4.55 (s, I , CHCMe3), 9.10 (s, 9, CHCMe3). 25. Preparation of TaCp[C(Ph)C(Ph)CHCMe3lClz.TaCp(CHCMe3)C12 (0.50 g) and 0.23 g of diphenylacetylene were stirred in 50 mL of toluene for 12 h. The toluene was concentrated to -5 mL in vacuo and 20 mL of pentane was added. Blue, crystalline TaCp[C(Ph)C(Ph)CHCMe3]CI2 (0.53 g, 72%) was isolated by filtration. Anal. Calcd for TaC24H2sCl2: C, 50.99; H, 4.45; CI, 12.54. Found: C, 50.81; H,4.54;CI, 12.44. 'H NMR (T,C6D6): 2.67-3.71 (m,-IO (olefinic proton buried in phenyl region, even at 270 MHz), Ph), 4.41 (s, 5, Cp), 8.69 ( s , 9, CHCMe3). I3cN M R (ppm, C6D6, ' H gated decoupled): 259 (s, Ta=CPh), 151 (s, Clpso),142 (s, C'lpso),137 (d, CHCMe3, ~ J C H= 152 Hz), 131 (s, =CPh), 131-126 (other Ph carbons), 1 I O (d, Cp, ~ J C H = 179 Hz), 36.1 (s, CHCMe3), 32.1 (q, CHCMe3, ' J C H = 126 Hz). 26. Reaction of TaCp[C(Ph)C(Ph)CHCMe3]Clz with CzH4. A solution of TaCp[C(Ph)C(Ph)CHCMe3]Cl2 (0.28 g) in I O mL of pentane was stirred under 50 psi of C2H4 for 12 h. Orange CpC12TaCH2CH2CH2CH2 (0.13 g, 72%) was isolated by filtration and identified by comparison of its ' H NMR spectrum with that of an authentic ample.^ The filtrate was concentrated to -3 mL and passed through a column of activated alumina. The residual pentane was removed in vacuo and 0.1 1 g of colorless, liquid CH2=CHCH(Ph)C(Ph)=CHCMe3 was isolated (79%). IH N M R (T,C6D6): 2.78-3.29 (m, IO, Ph), 3.92 (ddd, I , =CH-, J = 7, IO, and 18 Hz), 4.47 (s, I , =CHCMe3), 5.02 (m, 2, =CH2), 5.96 (d, I , -CH(Ph)-, J = 7 Hz), 9.13 (s, 9, =CHCMe3). I3C N M R (ppm, CDC13, IH off-resonance decoupled): 141.9, 141.5, 140.5 (s, Clpso,Clipso, -C(Ph)=), 140.I , 139.2 (d, =CH, =CHCMe3), 130-126 (other phenyl carbons), 115.7 (dd, =CH2), 60.0 (d, -CH(Ph)-), 33.5 (s,CHCMe3), 31.4 (q, CHCMe3). Decomposition Study of TaCp(CRzCMe3)zCIz(R = H or D). A solution of 25 mg of TaCp(CH2CMe3)2C12 and 150 mg of TaCp(CD2CMe3)2C12in 1 mL of chloroform was prepared in the dark and left there for 0.5 h. All volatiles were then removed in vacuo and analyzed by GC/mass spectroscopy for deuterated neopentanes as described previo~sly:'~ 41% dj, 7% d2, 2% d l , and 50% do. Another 1 mL of chloroform was added to the residue and the solution left for 5 h in the dark. The volatiles were again analyzed for deuterated neopentanes by GC/MS; 61% d3, 16% d2, 5% d l , and 18% do. A 250-mg sample of TaCp(CD2CMe3)2Cl2 was dissolved in 5.0 mL of benzene and the solution was heated at 60 "C for 4 h. The deuterated neopentanes (by GC/MS) were 94% d 3 and 6% d 2 ,

Acknowledgment. We thank the National Science Foundation for support (CHE76-07140), the Francis N. Bitter National Magnet Laboratory for use of their high field NMR facilities, and G . M. Whitesides for use of his G C / M S facilities. References and Notes (1)

Part 12: Clark, D. N.; Schrock. R . 6774.

R . J. Am.

Chem. SOC. 1978, 100.

(2) (a) NSF Predoctoral Fellow, 1975-1978.(b) Camille and Henry Dreyfus Teacher-Scholar, 1978-. (3)(a) Schrock, R. R.; Fellmann, J. D. J. Am. Chem. SOC. 1978, 100, 3359. (b) Schrock, R. R. bid. 1974, 96,6796. (4)TaCp(CHCMes)CI?and NbCp(CHCMe3)CI2have been reported in a pre-

liminary communication.5

(5) McLain. S.J.; Wood, C. D.; Schrock, R . R . J. Am. Chem. SOC.1977, 99, 3519.

(6) Schrock, R. R.; Messerle. L. W.; Wood, C. D.; Guggenberger,L. J. J. Am. Chem. SOC. 1978, 100,3793. (7)Churchill,M. R.; Youngs, W. J. horg. Chem. 1979, 18, 171. (8)(a) McLain, S.J. Unpublished results. (b) This complex is formed in poor

yield when TaCp"(CH2Ph)&I2decomposes thermally or photochemically: Messerle. L. W. Unpublished results. (9)The dramatic effect of substituting two 775-C5H5 with two v5-C5Me5rings in manganocene has been attributed to electron donation by the methyl groups: Smart, J. C.; Robbins, T. L. J. Am. Chem. Soc. 1978, 100. 3936.

3222

Journal of the American Chemical Society

(10) Schrock, R. R.; Sharp, P. R. J. Am. Chem. SOC.1978, 100, 2389. (1 1) Rupprecht, G. A. Unpublished results. (12) A similar type of reaction was reported between (CO)&rC(OMe)R (R = Me or Ph) and Et2NCECR' to give complexes of the type (CO),CrC(NEt,)C(R'~C(OMeXR): Dotz. K. H.; Kreiter, C. G. J. Organomet. Chem. 1975, 99, 309. Chem. Ber. 1976, 109, 2026. (13) Churchill, M. R.; Hollander, F. J. lnorg. Chem. 1978, 17, 1957. (14) Guggenberger. L. J.; Schrock, R. R. J. Am. Chem. SOC. 1975, 97, 6578. (15) Schultz, A. J.; Williams, J. M.; Schrock, R. R.; Rupprecht, G. A,; Fellmann, J. D. J. Am. Chem. SOC., 1979, 101, 1593. (16) (a) We thank Galen Stuckv for Dermission to ouote these results Drior to publication. (b) Friedrich, P.;Besl, G.; Fischer,'E. 0.; Huttner, G. j , Organomet. Chem. 1977, 139, C68. (17) (a) Oiah, G. A., Ed. "Friedel-Crafts and Related Reactions", Vol. I; WileyInterscience: New York, 1963; pp 249, 253, 855. (b) Brown. H. C.; Holmes, R. R. J. Am. Chem. SOC. 1956. 78. 2173. I c l "it mav be said in summarv that with few exceptions the heat of complex formatidn of Lewis acid met& halides and stability of their complexes increases in the order F < CI < Br < I and this order therefore also gives an indication of the Lewis acid strength , . ." (ref 17a, p 868). Quantitative data supporting the contention that Nb is a poorer Lewis acid than Ta in this case is unavailable. However, it is well-known that a heavier main group element is a better Lewis acid than that above it (e.g., SbF5 > AsF5 > PF5; ref 17a, p 197)and that this ordering would hold in complexes of the group 5 transition metals in their higher oxidation states (3+ or 5+) would be no surprise. (a) McLain, S. J.,; Wood, C. D. Messerle, L. W.; Schrock, R. R.; Hollander, F. J.; Youngs, W. J.; Churchill, M. R. J. Am. Chem. SOC.1978, 100, 5962. (b) Guggenberger, L. J.; Schrock, R. R. ibid. 1975, 97, 2935. (a) Wade, K. "Electron Deficient Compounds", Appleton-Century-Crofts: New York, 1971. (b) Lipscomb, W. N. "Boron Hydrides", W. A. Benjamin: New York, 1963; p 172. (a)Seyferth. D.; Singh. G. J. Am. Chem. SOC.1965, 87,4156. (b) Seyferth, D.; Hughes, W. B.; Heeren. J. K. ibid. 1965, 87, 2847. (a) Schmidbaur, H.; Richter, W.; Wolf, W.; Kohler, F. J. Chem. Ber. 1975, 108, 2649. (b) Ostoja Starzewski, K. A,; Feigel, M. J. Organomet. Chem. 1975, 93, C20. (c)Albright, T. A,; Freeman, W. J.; Schweizer, E. E. J. Am. Chem. SOC. 1975, 97,940. 'JcH in [Ph3PCH2CMe3]+ (6 31.0 in CDCI3)was found to be 140 f 5 Hz and in Ph3P=CHCMe3 (6 22.2 in CGD6)155 f 2 Hz; courtesy of D. N. Clark. It appears that Ta(cH~cMe3)~(0CMe~) decomposes to Ta(CHCMe3)(CH&Me3)2(OCMe$) at roughly the same rate in benzene or chloroform:

/ 101:12 / June 6, 1979

Felimann, J. D. Preliminary unpublished results. (251 Lamert. M. F.: Patil. D. S.; Pedley, J. B. J. Chem. SOC.,Chem. Commun. 1975, 830. (26) Calvert, R. B.; Shapley, J. R.; Schultz, A. J.; Williams, J. M.; Suib, S. L.; Stucky, G. D. J. Am. Chem. SOC.1978, 100, 6240. 1978, 100, 7726. (27) Calvert. R. B.; Shapley, J. R. J. Am. Chem. SOC. (28) Determining k H l k D by comparing the rate of decomposition of T ~ C P ( C H ~ C M ~ ~with ) ~ Cthat I , of TaCp(CD2CMe3)2C12gave an answer (5 4 f 0.5) which was just within experimental of that determined by decomposing TaCp(CHDCMe3)2C12(7 f 1). The two should be identical and independent of the relative contribution of any thermodynamic isotope discrimination to the overall isotope effect. (29) (a) Davies, G. R.; Jarvis, J. A,; Kilbourn. B. T. Chem. Commun. 1971, 1511. (b) Bassi, I. W.; Allegra, G.; Scordamaglia, R.; Chioccola, G. J. Am. Chem. SOC. 1971, 93, 3787. (c) Davies, G. R.; Jarvis, J.; Kilbourn, 8.T.; Pioli. A. J. P. Chem. Cornmun. 1971,677. (30) (a) The extreme exam le of this type of interaction is the "benzallyl" found ~ HRoberts, ~ ) . ~ ~ J. ~ ,S.; ~ in [Pd(CH2C6H5)C1]23gband M O C ~ ( C O ) ~ ( C H ~ C (b) Klabunde, J. K. J. Am. Chem. SOC. 1977, 99, 2509. (c) King, R. B.; Fronzaglia, A. ibid. 1966, 88, 709. (d) Cotton, F. A,; LaPrade, M. D. ibid. 1968, 90, 5418. (31) (a)Trofimenko, S. Inorg. Chem. 1970, 9, 2493. (b)Cotton, F. A,; Day, V. W. J. Chem. Soc., Chem. Commun. 1974, 415. (c) Cotton, F. A,; Stanislowski, A. G., J. Am. Chem. SOC.1974, 96, 5074. (d) Cotton, F. A,; LaCour, T.; Stanislowski, A. G. ibid. 1974, 96, 754. (32) (a) Brookhart. M.; Whitesides, T. H.; Crockett, J. M. lnorg. Chem. 1976, 15, 1550. (b)Olah,G. A.;Liang.G.; Yu, S. H. J. Org. Chem. 1976, 41, 2227. (c) Ittel, S. D.; Van-Catledge, F. A,; Tolman, C. A,; Jesson, J. P. J. Am. Chem. SOC.1978, 100, 1317. (33) Williams, J. M.; Brown, R . K.,; Schultz, A. J.; Stucky, G. D.; Ittel, S. D. J. Am. Chern. SOC.1978, 100, 7407. (34) Cooper, N. J.; Green, M. L. H. J. Chem. SOC.,Chem. Commun. 1974,206, 761. (35) Andersen, R. A,; Jones, R. A,; Wilkinson. G. J. Chem. SOC.,Dalton Trans. 1978, 446. (36) Whitesides, G. M. Personal communication. (37) Schrock, R. R. J. Organomet. Chem. 1976, 122, 209. (38) Threlkel, R. S.; Bercaw, J. R. J. Organomet. Chem. 1977, 136, 1. (39) Fritz, J. S.; Shenk, G. A., Jr. in "Qualitative Analytical Chemistry", 2nd ed.; Allyn and Bacon: Boston, Mass., 1969; pp 205-208, 534-537. (40) White, D. C. Mikrochim. Acta 1961, 449. (41) Van Baalen, A,; Groenenboom, C. J.; de Liefde Meijer, H. J. J. Orpnomet. Chem. 1974, 74, 245. (42) Watson, S. C.; Eastham, J. F. J. Organomet. Chem. 1967, 9, 165. ~i

Structural Studies of Organic Silver Complexes in Dimethyl Sulfoxide by 13C and lo9AgNMR P. M. Henrichs,* S. Sheard, J. J. H. Ackerman, and G . E. Maciel* Contribution from the Research Laboratories, Eastman Kodak Company, Rochester, New York 14650, and the Department of Chemistry, Colorado State University, Fort Collins, Colorado 80523. Received October I O , 1978

Abstract: The effects on I3C and lo9Agchemical shifts of complexation of ethylenethiourea, thiazolidine-2-thione, 3-methylthiazolidine-2-thione, dipropylamine, pentamethylene sulfide, and 2-aminothiazole with silver nitrate in dimethyl sulfoxide solution have been examined. Although complexation is reflected via lo9Ag chemical shifts of hundreds of parts per million, as compared with only a few parts per million of ligand 13Cchemical shifts, both types of N M R data suggest that ligands which bind Ag(1) through sulfur form solutions containing a variety of complex species. By contrast, nitrogen ligands appear to form primarily 2: 1 silver-to-ligand complexes. The I3C chemical-shift changes in the unsaturated compounds are discussed in terms of contributions from resonance forms containing different degrees of C=S and C=N bonding in the silver complexes compared to the free ligands.

Introduction We have shown in an earlier report' that a combination of I3C and lo9AgNMR holds promise for determining various structural parameters of organic silver complexes in solution. However, experimental data were obtained only for thiourea and tetramethylthiourea. The apparent structural diversity of the silver-complex systems for these compounds precluded detailed conclusions. Toward a long-term goal of providing a data base which may allow more of the information contained 0002-7863/79/1501-3222$01.00/0

in the I3C and lo9Ag spectra of silver complexes to be extracted, we have extended our study to include additional compounds. The I3C spectra of silver complexes may be useful in providing insight into the effects of complexing of metal ions in general on the I3C chemical shifts of organic ligands. Diamagnetic metal complexes have already been studied. Metal complexes of olefinic and aromatic ligands2q3 and the shifts caused by metal complexation in ~ a r b o n y land ~ , ~thiocarbonyI63' compounds have also been studied. A few investigations 0 1979 American Chemical Society